From 9539bfecb160d1581bfaa4ff5da22364f8d855b6 Mon Sep 17 00:00:00 2001
From: pmonormal <pmon2648@daum.net>
Date: Thu, 11 Oct 2018 12:50:16 +0900
Subject: [PATCH] Update Frauud detection Kernel as Korean about 78%
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

2.2 케라스 신경망 파트를 마무리 하고 있습니다.
---
 .../Fraud detection analysis with NN.ipynb    | 2407 +++++++++++++----
 1 file changed, 1823 insertions(+), 584 deletions(-)

diff --git a/Korean/Fraud detection analysis with NN/Fraud detection analysis with NN.ipynb b/Korean/Fraud detection analysis with NN/Fraud detection analysis with NN.ipynb
index 0addaef..ece030e 100644
--- a/Korean/Fraud detection analysis with NN/Fraud detection analysis with NN.ipynb	
+++ b/Korean/Fraud detection analysis with NN/Fraud detection analysis with NN.ipynb	
@@ -1975,27 +1975,27 @@
        "  </thead>\n",
        "  <tbody>\n",
        "    <tr>\n",
-       "      <th>143333</th>\n",
+       "      <th>16863</th>\n",
        "      <td>578.28934</td>\n",
        "      <td>0.0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>17480</th>\n",
+       "      <th>241254</th>\n",
        "      <td>578.28934</td>\n",
        "      <td>0.0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>6641</th>\n",
+       "      <th>76929</th>\n",
        "      <td>578.28934</td>\n",
        "      <td>0.0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>12369</th>\n",
+       "      <th>235634</th>\n",
        "      <td>578.28934</td>\n",
        "      <td>0.0</td>\n",
        "    </tr>\n",
        "    <tr>\n",
-       "      <th>87354</th>\n",
+       "      <th>23308</th>\n",
        "      <td>578.28934</td>\n",
        "      <td>0.0</td>\n",
        "    </tr>\n",
@@ -2005,11 +2005,11 @@
       ],
       "text/plain": [
        "            Fraud  Normal\n",
-       "143333  578.28934     0.0\n",
-       "17480   578.28934     0.0\n",
-       "6641    578.28934     0.0\n",
-       "12369   578.28934     0.0\n",
-       "87354   578.28934     0.0"
+       "16863   578.28934     0.0\n",
+       "241254  578.28934     0.0\n",
+       "76929   578.28934     0.0\n",
+       "235634  578.28934     0.0\n",
+       "23308   578.28934     0.0"
       ]
      },
      "execution_count": 30,
@@ -2145,240 +2145,240 @@
        "    </tr>\n",
        "    <tr>\n",
        "      <th>mean</th>\n",
-       "      <td>0.001775</td>\n",
-       "      <td>0.004433</td>\n",
-       "      <td>0.001241</td>\n",
-       "      <td>-0.002806</td>\n",
-       "      <td>-0.005931</td>\n",
-       "      <td>-0.001400</td>\n",
-       "      <td>-0.002919</td>\n",
-       "      <td>0.006406</td>\n",
-       "      <td>-0.006651</td>\n",
-       "      <td>-0.002855</td>\n",
-       "      <td>0.004204</td>\n",
-       "      <td>-0.002575</td>\n",
-       "      <td>0.003172</td>\n",
-       "      <td>0.000578</td>\n",
-       "      <td>-0.006489</td>\n",
-       "      <td>-0.006705</td>\n",
-       "      <td>-0.003444</td>\n",
-       "      <td>-0.000166</td>\n",
-       "      <td>0.004815</td>\n",
-       "      <td>-0.002209</td>\n",
-       "      <td>-0.001280</td>\n",
-       "      <td>-0.000847</td>\n",
-       "      <td>0.005033</td>\n",
-       "      <td>-0.005521</td>\n",
-       "      <td>0.000037</td>\n",
-       "      <td>-0.003663</td>\n",
-       "      <td>-0.000076</td>\n",
-       "      <td>0.004578</td>\n",
-       "      <td>0.004613</td>\n",
-       "      <td>0.000661</td>\n",
-       "      <td>-0.006340</td>\n",
+       "      <td>-0.001224</td>\n",
+       "      <td>-0.001042</td>\n",
+       "      <td>-0.002830</td>\n",
+       "      <td>-0.005544</td>\n",
+       "      <td>0.005997</td>\n",
+       "      <td>0.000293</td>\n",
+       "      <td>0.001331</td>\n",
+       "      <td>-0.005805</td>\n",
+       "      <td>-0.004224</td>\n",
+       "      <td>0.003585</td>\n",
+       "      <td>-0.003315</td>\n",
+       "      <td>0.006343</td>\n",
+       "      <td>0.014670</td>\n",
+       "      <td>-0.005487</td>\n",
+       "      <td>0.004601</td>\n",
+       "      <td>-0.002224</td>\n",
+       "      <td>-0.000504</td>\n",
+       "      <td>-0.001790</td>\n",
+       "      <td>-0.006839</td>\n",
+       "      <td>0.001810</td>\n",
+       "      <td>-0.002665</td>\n",
+       "      <td>-0.002930</td>\n",
+       "      <td>0.001825</td>\n",
+       "      <td>0.000312</td>\n",
+       "      <td>0.002653</td>\n",
+       "      <td>-0.001334</td>\n",
+       "      <td>0.001097</td>\n",
+       "      <td>0.000116</td>\n",
+       "      <td>0.000894</td>\n",
+       "      <td>0.001173</td>\n",
+       "      <td>-0.000229</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>std</th>\n",
-       "      <td>1.001101</td>\n",
-       "      <td>0.994830</td>\n",
-       "      <td>0.996210</td>\n",
-       "      <td>0.986656</td>\n",
-       "      <td>0.997242</td>\n",
-       "      <td>0.977614</td>\n",
-       "      <td>0.990565</td>\n",
-       "      <td>0.969458</td>\n",
-       "      <td>1.003326</td>\n",
-       "      <td>1.001119</td>\n",
-       "      <td>0.994826</td>\n",
-       "      <td>0.993596</td>\n",
-       "      <td>0.989928</td>\n",
-       "      <td>1.002852</td>\n",
-       "      <td>0.998493</td>\n",
        "      <td>1.001920</td>\n",
-       "      <td>0.997151</td>\n",
-       "      <td>0.993978</td>\n",
-       "      <td>1.000924</td>\n",
-       "      <td>0.996003</td>\n",
-       "      <td>0.992029</td>\n",
-       "      <td>1.008943</td>\n",
-       "      <td>1.002597</td>\n",
-       "      <td>0.982256</td>\n",
-       "      <td>0.995826</td>\n",
-       "      <td>1.004384</td>\n",
-       "      <td>1.002220</td>\n",
-       "      <td>0.984900</td>\n",
-       "      <td>0.990007</td>\n",
-       "      <td>0.982990</td>\n",
-       "      <td>0.928776</td>\n",
+       "      <td>1.020089</td>\n",
+       "      <td>1.018808</td>\n",
+       "      <td>1.011348</td>\n",
+       "      <td>0.998416</td>\n",
+       "      <td>0.999303</td>\n",
+       "      <td>0.998207</td>\n",
+       "      <td>1.018658</td>\n",
+       "      <td>1.088381</td>\n",
+       "      <td>0.991733</td>\n",
+       "      <td>0.990113</td>\n",
+       "      <td>0.997706</td>\n",
+       "      <td>0.996597</td>\n",
+       "      <td>1.001690</td>\n",
+       "      <td>0.990480</td>\n",
+       "      <td>0.996792</td>\n",
+       "      <td>0.996392</td>\n",
+       "      <td>0.991946</td>\n",
+       "      <td>0.997001</td>\n",
+       "      <td>1.001586</td>\n",
+       "      <td>1.009827</td>\n",
+       "      <td>1.068501</td>\n",
+       "      <td>1.009975</td>\n",
+       "      <td>1.072639</td>\n",
+       "      <td>1.003264</td>\n",
+       "      <td>1.003527</td>\n",
+       "      <td>1.003375</td>\n",
+       "      <td>1.010746</td>\n",
+       "      <td>1.039948</td>\n",
+       "      <td>0.985931</td>\n",
+       "      <td>0.997525</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>min</th>\n",
        "      <td>-1.996369</td>\n",
-       "      <td>-21.406457</td>\n",
-       "      <td>-36.616183</td>\n",
-       "      <td>-21.741294</td>\n",
-       "      <td>-3.955598</td>\n",
-       "      <td>-23.251009</td>\n",
-       "      <td>-15.009219</td>\n",
-       "      <td>-25.218244</td>\n",
-       "      <td>-42.215404</td>\n",
-       "      <td>-8.613050</td>\n",
-       "      <td>-20.376630</td>\n",
-       "      <td>-4.587901</td>\n",
-       "      <td>-17.783345</td>\n",
-       "      <td>-4.027674</td>\n",
-       "      <td>-20.044245</td>\n",
-       "      <td>-4.397796</td>\n",
-       "      <td>-15.478720</td>\n",
-       "      <td>-29.626400</td>\n",
-       "      <td>-11.081002</td>\n",
-       "      <td>-8.523283</td>\n",
+       "      <td>-20.661780</td>\n",
+       "      <td>-28.722479</td>\n",
+       "      <td>-22.213271</td>\n",
+       "      <td>-3.717201</td>\n",
+       "      <td>-21.540062</td>\n",
+       "      <td>-15.052954</td>\n",
+       "      <td>-33.551864</td>\n",
+       "      <td>-61.302416</td>\n",
+       "      <td>-12.227994</td>\n",
+       "      <td>-22.411893</td>\n",
+       "      <td>-4.363472</td>\n",
+       "      <td>-17.242432</td>\n",
+       "      <td>-5.819382</td>\n",
+       "      <td>-19.292570</td>\n",
+       "      <td>-4.584886</td>\n",
+       "      <td>-14.141290</td>\n",
+       "      <td>-26.540290</td>\n",
+       "      <td>-10.846039</td>\n",
+       "      <td>-6.059567</td>\n",
        "      <td>-29.624863</td>\n",
-       "      <td>-31.162149</td>\n",
-       "      <td>-13.089986</td>\n",
-       "      <td>-36.151218</td>\n",
-       "      <td>-4.566222</td>\n",
-       "      <td>-13.584545</td>\n",
-       "      <td>-5.401088</td>\n",
+       "      <td>-47.418984</td>\n",
+       "      <td>-12.246104</td>\n",
+       "      <td>-58.716303</td>\n",
+       "      <td>-4.584519</td>\n",
+       "      <td>-16.683278</td>\n",
+       "      <td>-3.847473</td>\n",
        "      <td>-24.515479</td>\n",
-       "      <td>-24.804228</td>\n",
+       "      <td>-25.260665</td>\n",
        "      <td>-0.353229</td>\n",
        "      <td>-0.046062</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>25%</th>\n",
-       "      <td>-0.860612</td>\n",
-       "      <td>-0.462792</td>\n",
-       "      <td>-0.359678</td>\n",
-       "      <td>-0.591207</td>\n",
-       "      <td>-0.601024</td>\n",
-       "      <td>-0.506230</td>\n",
-       "      <td>-0.580421</td>\n",
-       "      <td>-0.444144</td>\n",
-       "      <td>-0.176127</td>\n",
-       "      <td>-0.589594</td>\n",
-       "      <td>-0.490031</td>\n",
-       "      <td>-0.745758</td>\n",
-       "      <td>-0.406042</td>\n",
-       "      <td>-0.655290</td>\n",
-       "      <td>-0.451371</td>\n",
-       "      <td>-0.643494</td>\n",
-       "      <td>-0.540564</td>\n",
-       "      <td>-0.571829</td>\n",
-       "      <td>-0.589950</td>\n",
-       "      <td>-0.563178</td>\n",
-       "      <td>-0.274852</td>\n",
-       "      <td>-0.310767</td>\n",
-       "      <td>-0.747075</td>\n",
-       "      <td>-0.258595</td>\n",
-       "      <td>-0.588752</td>\n",
-       "      <td>-0.612822</td>\n",
-       "      <td>-0.679188</td>\n",
-       "      <td>-0.174095</td>\n",
-       "      <td>-0.159881</td>\n",
+       "      <td>-0.858169</td>\n",
+       "      <td>-0.470222</td>\n",
+       "      <td>-0.358188</td>\n",
+       "      <td>-0.589043</td>\n",
+       "      <td>-0.592879</td>\n",
+       "      <td>-0.497053</td>\n",
+       "      <td>-0.575976</td>\n",
+       "      <td>-0.445508</td>\n",
+       "      <td>-0.175328</td>\n",
+       "      <td>-0.575156</td>\n",
+       "      <td>-0.487916</td>\n",
+       "      <td>-0.740248</td>\n",
+       "      <td>-0.391293</td>\n",
+       "      <td>-0.656250</td>\n",
+       "      <td>-0.439695</td>\n",
+       "      <td>-0.634752</td>\n",
+       "      <td>-0.535462</td>\n",
+       "      <td>-0.571714</td>\n",
+       "      <td>-0.596332</td>\n",
+       "      <td>-0.557152</td>\n",
+       "      <td>-0.272606</td>\n",
+       "      <td>-0.309250</td>\n",
+       "      <td>-0.744022</td>\n",
+       "      <td>-0.257902</td>\n",
+       "      <td>-0.579922</td>\n",
+       "      <td>-0.605941</td>\n",
+       "      <td>-0.679235</td>\n",
+       "      <td>-0.175332</td>\n",
+       "      <td>-0.159296</td>\n",
        "      <td>-0.331279</td>\n",
        "      <td>-0.046062</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>50%</th>\n",
-       "      <td>-0.205438</td>\n",
-       "      <td>0.009624</td>\n",
-       "      <td>0.041337</td>\n",
-       "      <td>0.114507</td>\n",
-       "      <td>-0.018562</td>\n",
-       "      <td>-0.039359</td>\n",
-       "      <td>-0.208009</td>\n",
-       "      <td>0.034818</td>\n",
-       "      <td>0.015724</td>\n",
-       "      <td>-0.047991</td>\n",
-       "      <td>-0.084941</td>\n",
-       "      <td>-0.033111</td>\n",
-       "      <td>0.140522</td>\n",
-       "      <td>-0.010926</td>\n",
-       "      <td>0.049060</td>\n",
-       "      <td>0.045656</td>\n",
-       "      <td>0.072575</td>\n",
-       "      <td>-0.081757</td>\n",
-       "      <td>-0.001062</td>\n",
-       "      <td>-0.000870</td>\n",
-       "      <td>-0.081319</td>\n",
-       "      <td>-0.039553</td>\n",
-       "      <td>0.015440</td>\n",
-       "      <td>-0.019763</td>\n",
-       "      <td>0.066587</td>\n",
-       "      <td>0.028921</td>\n",
-       "      <td>-0.108500</td>\n",
-       "      <td>0.004913</td>\n",
-       "      <td>0.035269</td>\n",
-       "      <td>-0.264951</td>\n",
+       "      <td>-0.214577</td>\n",
+       "      <td>0.013992</td>\n",
+       "      <td>0.035383</td>\n",
+       "      <td>0.112973</td>\n",
+       "      <td>-0.000758</td>\n",
+       "      <td>-0.042363</td>\n",
+       "      <td>-0.203824</td>\n",
+       "      <td>0.030520</td>\n",
+       "      <td>0.020143</td>\n",
+       "      <td>-0.039460</td>\n",
+       "      <td>-0.085708</td>\n",
+       "      <td>-0.024836</td>\n",
+       "      <td>0.153613</td>\n",
+       "      <td>-0.018689</td>\n",
+       "      <td>0.055610</td>\n",
+       "      <td>0.047554</td>\n",
+       "      <td>0.077496</td>\n",
+       "      <td>-0.079770</td>\n",
+       "      <td>-0.010362</td>\n",
+       "      <td>0.005983</td>\n",
+       "      <td>-0.080698</td>\n",
+       "      <td>-0.041409</td>\n",
+       "      <td>0.007294</td>\n",
+       "      <td>-0.017332</td>\n",
+       "      <td>0.072023</td>\n",
+       "      <td>0.028636</td>\n",
+       "      <td>-0.115717</td>\n",
+       "      <td>0.003963</td>\n",
+       "      <td>0.032194</td>\n",
+       "      <td>-0.263912</td>\n",
        "      <td>-0.046062</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>75%</th>\n",
-       "      <td>0.937963</td>\n",
-       "      <td>0.671790</td>\n",
-       "      <td>0.486854</td>\n",
-       "      <td>0.671008</td>\n",
-       "      <td>0.521349</td>\n",
-       "      <td>0.443249</td>\n",
-       "      <td>0.298778</td>\n",
-       "      <td>0.468235</td>\n",
-       "      <td>0.270448</td>\n",
-       "      <td>0.538766</td>\n",
-       "      <td>0.422601</td>\n",
-       "      <td>0.719480</td>\n",
-       "      <td>0.622925</td>\n",
-       "      <td>0.673880</td>\n",
-       "      <td>0.511166</td>\n",
-       "      <td>0.703773</td>\n",
-       "      <td>0.591952</td>\n",
-       "      <td>0.474241</td>\n",
-       "      <td>0.600049</td>\n",
-       "      <td>0.563023</td>\n",
-       "      <td>0.172138</td>\n",
-       "      <td>0.254958</td>\n",
-       "      <td>0.739301</td>\n",
-       "      <td>0.235829</td>\n",
-       "      <td>0.730778</td>\n",
-       "      <td>0.673369</td>\n",
-       "      <td>0.502075</td>\n",
-       "      <td>0.227578</td>\n",
-       "      <td>0.240531</td>\n",
-       "      <td>-0.041379</td>\n",
+       "      <td>0.939522</td>\n",
+       "      <td>0.672490</td>\n",
+       "      <td>0.487033</td>\n",
+       "      <td>0.677709</td>\n",
+       "      <td>0.534019</td>\n",
+       "      <td>0.439335</td>\n",
+       "      <td>0.301598</td>\n",
+       "      <td>0.458236</td>\n",
+       "      <td>0.275742</td>\n",
+       "      <td>0.550495</td>\n",
+       "      <td>0.403495</td>\n",
+       "      <td>0.730971</td>\n",
+       "      <td>0.625753</td>\n",
+       "      <td>0.662847</td>\n",
+       "      <td>0.513317</td>\n",
+       "      <td>0.699394</td>\n",
+       "      <td>0.593329</td>\n",
+       "      <td>0.468597</td>\n",
+       "      <td>0.592829</td>\n",
+       "      <td>0.564164</td>\n",
+       "      <td>0.173084</td>\n",
+       "      <td>0.252904</td>\n",
+       "      <td>0.733052</td>\n",
+       "      <td>0.236988</td>\n",
+       "      <td>0.729891</td>\n",
+       "      <td>0.668831</td>\n",
+       "      <td>0.496408</td>\n",
+       "      <td>0.227238</td>\n",
+       "      <td>0.231837</td>\n",
+       "      <td>-0.039260</td>\n",
        "      <td>-0.046062</td>\n",
        "    </tr>\n",
        "    <tr>\n",
        "      <th>max</th>\n",
-       "      <td>1.641507</td>\n",
-       "      <td>1.253349</td>\n",
-       "      <td>11.011647</td>\n",
-       "      <td>2.705162</td>\n",
-       "      <td>11.918722</td>\n",
-       "      <td>20.838790</td>\n",
-       "      <td>16.057594</td>\n",
-       "      <td>27.728845</td>\n",
-       "      <td>16.049132</td>\n",
-       "      <td>8.117551</td>\n",
-       "      <td>14.080677</td>\n",
-       "      <td>11.775017</td>\n",
-       "      <td>4.476496</td>\n",
-       "      <td>4.486616</td>\n",
-       "      <td>8.089541</td>\n",
-       "      <td>6.249731</td>\n",
-       "      <td>7.352670</td>\n",
-       "      <td>9.293592</td>\n",
-       "      <td>5.622204</td>\n",
-       "      <td>5.792269</td>\n",
-       "      <td>49.443471</td>\n",
+       "      <td>1.641929</td>\n",
+       "      <td>1.229413</td>\n",
+       "      <td>13.000116</td>\n",
+       "      <td>2.657487</td>\n",
+       "      <td>8.568820</td>\n",
+       "      <td>21.022418</td>\n",
+       "      <td>15.993545</td>\n",
+       "      <td>25.484930</td>\n",
+       "      <td>15.697933</td>\n",
+       "      <td>9.388014</td>\n",
+       "      <td>12.531076</td>\n",
+       "      <td>11.432405</td>\n",
+       "      <td>7.854665</td>\n",
+       "      <td>3.923102</td>\n",
+       "      <td>10.981446</td>\n",
+       "      <td>6.211952</td>\n",
+       "      <td>6.528161</td>\n",
+       "      <td>8.592238</td>\n",
+       "      <td>4.522099</td>\n",
+       "      <td>5.897892</td>\n",
+       "      <td>34.033648</td>\n",
        "      <td>37.034649</td>\n",
-       "      <td>9.988628</td>\n",
-       "      <td>27.700472</td>\n",
-       "      <td>6.628356</td>\n",
+       "      <td>14.473016</td>\n",
+       "      <td>30.430985</td>\n",
+       "      <td>6.642261</td>\n",
        "      <td>14.425293</td>\n",
-       "      <td>6.468519</td>\n",
-       "      <td>26.033296</td>\n",
+       "      <td>7.181774</td>\n",
+       "      <td>21.576490</td>\n",
        "      <td>68.528383</td>\n",
-       "      <td>75.250448</td>\n",
+       "      <td>51.265692</td>\n",
        "      <td>21.709793</td>\n",
        "    </tr>\n",
        "  </tbody>\n",
@@ -2388,68 +2388,68 @@
       "text/plain": [
        "               Time            V1            V2            V3            V4  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean       0.001775      0.004433      0.001241     -0.002806     -0.005931   \n",
-       "std        1.001101      0.994830      0.996210      0.986656      0.997242   \n",
-       "min       -1.996369    -21.406457    -36.616183    -21.741294     -3.955598   \n",
-       "25%       -0.860612     -0.462792     -0.359678     -0.591207     -0.601024   \n",
-       "50%       -0.205438      0.009624      0.041337      0.114507     -0.018562   \n",
-       "75%        0.937963      0.671790      0.486854      0.671008      0.521349   \n",
-       "max        1.641507      1.253349     11.011647      2.705162     11.918722   \n",
+       "mean      -0.001224     -0.001042     -0.002830     -0.005544      0.005997   \n",
+       "std        1.001920      1.020089      1.018808      1.011348      0.998416   \n",
+       "min       -1.996369    -20.661780    -28.722479    -22.213271     -3.717201   \n",
+       "25%       -0.858169     -0.470222     -0.358188     -0.589043     -0.592879   \n",
+       "50%       -0.214577      0.013992      0.035383      0.112973     -0.000758   \n",
+       "75%        0.939522      0.672490      0.487033      0.677709      0.534019   \n",
+       "max        1.641929      1.229413     13.000116      2.657487      8.568820   \n",
        "\n",
        "                 V5            V6            V7            V8            V9  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean      -0.001400     -0.002919      0.006406     -0.006651     -0.002855   \n",
-       "std        0.977614      0.990565      0.969458      1.003326      1.001119   \n",
-       "min      -23.251009    -15.009219    -25.218244    -42.215404     -8.613050   \n",
-       "25%       -0.506230     -0.580421     -0.444144     -0.176127     -0.589594   \n",
-       "50%       -0.039359     -0.208009      0.034818      0.015724     -0.047991   \n",
-       "75%        0.443249      0.298778      0.468235      0.270448      0.538766   \n",
-       "max       20.838790     16.057594     27.728845     16.049132      8.117551   \n",
+       "mean       0.000293      0.001331     -0.005805     -0.004224      0.003585   \n",
+       "std        0.999303      0.998207      1.018658      1.088381      0.991733   \n",
+       "min      -21.540062    -15.052954    -33.551864    -61.302416    -12.227994   \n",
+       "25%       -0.497053     -0.575976     -0.445508     -0.175328     -0.575156   \n",
+       "50%       -0.042363     -0.203824      0.030520      0.020143     -0.039460   \n",
+       "75%        0.439335      0.301598      0.458236      0.275742      0.550495   \n",
+       "max       21.022418     15.993545     25.484930     15.697933      9.388014   \n",
        "\n",
        "                V10           V11           V12           V13           V14  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean       0.004204     -0.002575      0.003172      0.000578     -0.006489   \n",
-       "std        0.994826      0.993596      0.989928      1.002852      0.998493   \n",
-       "min      -20.376630     -4.587901    -17.783345     -4.027674    -20.044245   \n",
-       "25%       -0.490031     -0.745758     -0.406042     -0.655290     -0.451371   \n",
-       "50%       -0.084941     -0.033111      0.140522     -0.010926      0.049060   \n",
-       "75%        0.422601      0.719480      0.622925      0.673880      0.511166   \n",
-       "max       14.080677     11.775017      4.476496      4.486616      8.089541   \n",
+       "mean      -0.003315      0.006343      0.014670     -0.005487      0.004601   \n",
+       "std        0.990113      0.997706      0.996597      1.001690      0.990480   \n",
+       "min      -22.411893     -4.363472    -17.242432     -5.819382    -19.292570   \n",
+       "25%       -0.487916     -0.740248     -0.391293     -0.656250     -0.439695   \n",
+       "50%       -0.085708     -0.024836      0.153613     -0.018689      0.055610   \n",
+       "75%        0.403495      0.730971      0.625753      0.662847      0.513317   \n",
+       "max       12.531076     11.432405      7.854665      3.923102     10.981446   \n",
        "\n",
        "                V15           V16           V17           V18           V19  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean      -0.006705     -0.003444     -0.000166      0.004815     -0.002209   \n",
-       "std        1.001920      0.997151      0.993978      1.000924      0.996003   \n",
-       "min       -4.397796    -15.478720    -29.626400    -11.081002     -8.523283   \n",
-       "25%       -0.643494     -0.540564     -0.571829     -0.589950     -0.563178   \n",
-       "50%        0.045656      0.072575     -0.081757     -0.001062     -0.000870   \n",
-       "75%        0.703773      0.591952      0.474241      0.600049      0.563023   \n",
-       "max        6.249731      7.352670      9.293592      5.622204      5.792269   \n",
+       "mean      -0.002224     -0.000504     -0.001790     -0.006839      0.001810   \n",
+       "std        0.996792      0.996392      0.991946      0.997001      1.001586   \n",
+       "min       -4.584886    -14.141290    -26.540290    -10.846039     -6.059567   \n",
+       "25%       -0.634752     -0.535462     -0.571714     -0.596332     -0.557152   \n",
+       "50%        0.047554      0.077496     -0.079770     -0.010362      0.005983   \n",
+       "75%        0.699394      0.593329      0.468597      0.592829      0.564164   \n",
+       "max        6.211952      6.528161      8.592238      4.522099      5.897892   \n",
        "\n",
        "                V20           V21           V22           V23           V24  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean      -0.001280     -0.000847      0.005033     -0.005521      0.000037   \n",
-       "std        0.992029      1.008943      1.002597      0.982256      0.995826   \n",
-       "min      -29.624863    -31.162149    -13.089986    -36.151218     -4.566222   \n",
-       "25%       -0.274852     -0.310767     -0.747075     -0.258595     -0.588752   \n",
-       "50%       -0.081319     -0.039553      0.015440     -0.019763      0.066587   \n",
-       "75%        0.172138      0.254958      0.739301      0.235829      0.730778   \n",
-       "max       49.443471     37.034649      9.988628     27.700472      6.628356   \n",
+       "mean      -0.002665     -0.002930      0.001825      0.000312      0.002653   \n",
+       "std        1.009827      1.068501      1.009975      1.072639      1.003264   \n",
+       "min      -29.624863    -47.418984    -12.246104    -58.716303     -4.584519   \n",
+       "25%       -0.272606     -0.309250     -0.744022     -0.257902     -0.579922   \n",
+       "50%       -0.080698     -0.041409      0.007294     -0.017332      0.072023   \n",
+       "75%        0.173084      0.252904      0.733052      0.236988      0.729891   \n",
+       "max       34.033648     37.034649     14.473016     30.430985      6.642261   \n",
        "\n",
        "                V25           V26           V27           V28        Amount  \\\n",
        "count  56961.000000  56961.000000  56961.000000  56961.000000  56961.000000   \n",
-       "mean      -0.003663     -0.000076      0.004578      0.004613      0.000661   \n",
-       "std        1.004384      1.002220      0.984900      0.990007      0.982990   \n",
-       "min      -13.584545     -5.401088    -24.515479    -24.804228     -0.353229   \n",
-       "25%       -0.612822     -0.679188     -0.174095     -0.159881     -0.331279   \n",
-       "50%        0.028921     -0.108500      0.004913      0.035269     -0.264951   \n",
-       "75%        0.673369      0.502075      0.227578      0.240531     -0.041379   \n",
-       "max       14.425293      6.468519     26.033296     68.528383     75.250448   \n",
+       "mean      -0.001334      0.001097      0.000116      0.000894      0.001173   \n",
+       "std        1.003527      1.003375      1.010746      1.039948      0.985931   \n",
+       "min      -16.683278     -3.847473    -24.515479    -25.260665     -0.353229   \n",
+       "25%       -0.605941     -0.679235     -0.175332     -0.159296     -0.331279   \n",
+       "50%        0.028636     -0.115717      0.003963      0.032194     -0.263912   \n",
+       "75%        0.668831      0.496408      0.227238      0.231837     -0.039260   \n",
+       "max       14.425293      7.181774     21.576490     68.528383     51.265692   \n",
        "\n",
        "       Amount_max_fraud  \n",
        "count      56961.000000  \n",
-       "mean          -0.006340  \n",
-       "std            0.928776  \n",
+       "mean          -0.000229  \n",
+       "std            0.997525  \n",
        "min           -0.046062  \n",
        "25%           -0.046062  \n",
        "50%           -0.046062  \n",
@@ -2645,7 +2645,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 40,
+   "execution_count": 39,
    "metadata": {
     "_cell_guid": "3f0b796a-8454-4cc9-b5ec-3075b9565bc3",
     "_uuid": "825805efc899b013c6bf7eaebdf10a68aee89d1a",
@@ -2656,27 +2656,27 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Epoch: 0 Acc = 0.93682 Cost = 72292.81250 Valid_Acc = 0.93725 Valid_Cost =  8597.14648\n",
-      "Epoch: 5 Acc = 0.93256 Cost = 59755.20703 Valid_Acc = 0.93174 Valid_Cost =  7915.83301\n",
-      "Epoch: 10 Acc = 0.93574 Cost = 53739.06250 Valid_Acc = 0.93490 Valid_Cost =  7608.52051\n",
-      "Epoch: 15 Acc = 0.95158 Cost = 42080.82031 Valid_Acc = 0.95004 Valid_Cost =  7097.75000\n",
-      "Epoch: 20 Acc = 0.96637 Cost = 31364.65234 Valid_Acc = 0.96545 Valid_Cost =  6864.34570\n",
-      "Epoch: 25 Acc = 0.98125 Cost = 21488.47266 Valid_Acc = 0.97978 Valid_Cost =  7460.70605\n",
-      "Epoch: 30 Acc = 0.98849 Cost = 14415.88770 Valid_Acc = 0.98785 Valid_Cost =  8910.00977\n",
-      "Epoch: 35 Acc = 0.98893 Cost = 13119.75000 Valid_Acc = 0.98820 Valid_Cost =  11685.54688\n",
-      "Epoch: 40 Acc = 0.98359 Cost = 12100.78711 Valid_Acc = 0.98227 Valid_Cost =  11338.14648\n",
-      "Epoch: 45 Acc = 0.98684 Cost = 10194.38477 Valid_Acc = 0.98617 Valid_Cost =  12333.08691\n",
-      "Epoch: 50 Acc = 0.95153 Cost = 30340.25000 Valid_Acc = 0.94993 Valid_Cost =  12908.94727\n",
-      "Epoch: 55 Acc = 0.99233 Cost = 6217.56738 Valid_Acc = 0.99122 Valid_Cost =  18185.65234\n",
-      "Epoch: 60 Acc = 0.98603 Cost = 10973.65039 Valid_Acc = 0.98490 Valid_Cost =  22597.33984\n",
-      "Epoch: 65 Acc = 0.99398 Cost = 5309.19678 Valid_Acc = 0.99333 Valid_Cost =  24412.04297\n",
-      "Epoch: 70 Acc = 0.99533 Cost = 4424.54004 Valid_Acc = 0.99438 Valid_Cost =  33113.28516\n",
-      "Epoch: 75 Acc = 0.98873 Cost = 9013.62500 Valid_Acc = 0.98718 Valid_Cost =  20797.70117\n",
-      "Epoch: 80 Acc = 0.99629 Cost = 3814.94922 Valid_Acc = 0.99551 Valid_Cost =  32202.07227\n",
-      "Epoch: 85 Acc = 0.99641 Cost = 3504.27686 Valid_Acc = 0.99547 Valid_Cost =  39097.82812\n",
-      "Epoch: 90 Acc = 0.99131 Cost = 6182.81982 Valid_Acc = 0.98978 Valid_Cost =  30556.54883\n",
-      "Epoch: 95 Acc = 0.97816 Cost = 17193.08203 Valid_Acc = 0.97647 Valid_Cost =  24212.60938\n",
-      "Epoch: 100 Acc = 0.99717 Cost = 2946.98804 Valid_Acc = 0.99614 Valid_Cost =  41408.02344\n",
+      "Epoch: 0 Acc = 0.98030 Cost = 70089.28125 Valid_Acc = 0.98034 Valid_Cost =  15254.80078\n",
+      "Epoch: 5 Acc = 0.98767 Cost = 58580.29297 Valid_Acc = 0.98750 Valid_Cost =  14394.41895\n",
+      "Epoch: 10 Acc = 0.98854 Cost = 48541.71875 Valid_Acc = 0.98883 Valid_Cost =  14806.13574\n",
+      "Epoch: 15 Acc = 0.98865 Cost = 39388.20703 Valid_Acc = 0.98894 Valid_Cost =  16517.65625\n",
+      "Epoch: 20 Acc = 0.99094 Cost = 31618.80859 Valid_Acc = 0.99122 Valid_Cost =  20146.30859\n",
+      "Epoch: 25 Acc = 0.99130 Cost = 22800.29883 Valid_Acc = 0.99136 Valid_Cost =  24841.26758\n",
+      "Epoch: 30 Acc = 0.99118 Cost = 18227.08789 Valid_Acc = 0.99143 Valid_Cost =  28714.44141\n",
+      "Epoch: 35 Acc = 0.99246 Cost = 11901.30176 Valid_Acc = 0.99305 Valid_Cost =  36806.41406\n",
+      "Epoch: 40 Acc = 0.99238 Cost = 14242.89941 Valid_Acc = 0.99238 Valid_Cost =  45021.91797\n",
+      "Epoch: 45 Acc = 0.99631 Cost = 9457.01465 Valid_Acc = 0.99586 Valid_Cost =  56201.80469\n",
+      "Epoch: 50 Acc = 0.99604 Cost = 10883.57422 Valid_Acc = 0.99582 Valid_Cost =  58257.39453\n",
+      "Epoch: 55 Acc = 0.99695 Cost = 6030.76562 Valid_Acc = 0.99666 Valid_Cost =  62535.69531\n",
+      "Epoch: 60 Acc = 0.99780 Cost = 6262.52979 Valid_Acc = 0.99758 Valid_Cost =  74222.21094\n",
+      "Epoch: 65 Acc = 0.99716 Cost = 4710.59668 Valid_Acc = 0.99673 Valid_Cost =  65601.07031\n",
+      "Epoch: 70 Acc = 0.99813 Cost = 4759.97021 Valid_Acc = 0.99726 Valid_Cost =  75777.33594\n",
+      "Epoch: 75 Acc = 0.99081 Cost = 6893.59326 Valid_Acc = 0.99119 Valid_Cost =  59490.25781\n",
+      "Epoch: 80 Acc = 0.99744 Cost = 2822.39600 Valid_Acc = 0.99695 Valid_Cost =  82329.80469\n",
+      "Epoch: 85 Acc = 0.99785 Cost = 2602.71289 Valid_Acc = 0.99740 Valid_Cost =  94100.67969\n",
+      "Epoch: 90 Acc = 0.99610 Cost = 3647.93262 Valid_Acc = 0.99582 Valid_Cost =  75909.85156\n",
+      "Epoch: 95 Acc = 0.99828 Cost = 3060.12622 Valid_Acc = 0.99775 Valid_Cost =  99243.14844\n",
+      "Epoch: 100 Acc = 0.99632 Cost = 3478.99805 Valid_Acc = 0.99558 Valid_Cost =  93043.17969\n",
       "\n",
       "Optimization Finished!\n",
       "\n",
@@ -2761,7 +2761,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 41,
+   "execution_count": 40,
    "metadata": {
     "_cell_guid": "c04813ad-3bbb-45c5-8098-06a78605d7b2",
     "_uuid": "2cc553c14ddcfcee1e3f833238da31fa18688c1e"
@@ -2769,7 +2769,7 @@
    "outputs": [
     {
      "data": {
-      "image/png": "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\n",
+      "image/png": "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\n",
       "text/plain": [
        "<Figure size 720x288 with 2 Axes>"
       ]
@@ -2807,7 +2807,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 42,
+   "execution_count": 41,
    "metadata": {
     "_cell_guid": "362ce01d-2582-47ea-9c04-3f2eb71dd87b",
     "_uuid": "a37daaf1fe266f2ee7840d08b9406bfa1b9ee841"
@@ -2818,7 +2818,7 @@
      "output_type": "stream",
      "text": [
       "INFO:tensorflow:Restoring parameters from ./model/model.ckpt\n",
-      "Testing Accuracy = 0.9968049\n"
+      "Testing Accuracy = 0.9961027\n"
      ]
     }
    ],
@@ -2841,27 +2841,29 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "32e42917-d3a8-4c4d-8eba-e21aeb743b70",
-    "_uuid": "09dbb6a1dbb178b242f9f2834f8deb390590c0f1"
-   },
+   "metadata": {},
    "source": [
-    "## 2.2 Train Neural Net with Keras\n",
+    "## 2.2 Keras 신경망 학습\n",
     "\n",
-    "Below I am implementing a similar NN with Keras starting from the original dataset and based on the understanding in EDA above.\n",
-    "Keras is high level framework (i.e. simpler codes) that actually runs on Tensorflow.\n",
-    "We shall evaluate NN model accuracy with Keras"
+    "아래의 코드는 원본 데이터와 EDA 이해를 바탕으로 Keras를 사용한 신경망(NN)입니다. Keras는 Tensorflow로 구성된 High Level 프레임워크입니다. (Higl Level 프레임워크는 코드를 간결하게 만들어, 코드 이해를 쉽게 만듭니다.) Keras로 신경망(NN) 모델의 정확도를 계산해 보겠습니다."
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 42,
    "metadata": {
     "_cell_guid": "83d06e55-189a-4acb-bb42-934accebec99",
-    "_uuid": "6918fa3e8198317d3b4e706e01366831ecc86c1e",
-    "collapsed": true
+    "_uuid": "6918fa3e8198317d3b4e706e01366831ecc86c1e"
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Using TensorFlow backend.\n"
+     ]
+    }
+   ],
    "source": [
     "import itertools\n",
     "from keras.models import Sequential\n",
@@ -2871,317 +2873,1556 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 43,
    "metadata": {
     "_cell_guid": "0e090edc-f544-41b6-8035-bf61c9e86205",
-    "_uuid": "d0bdc4d26b0a9f23c9ed7df460d57e2f489ef164",
-    "collapsed": true
+    "_uuid": "d0bdc4d26b0a9f23c9ed7df460d57e2f489ef164"
    },
    "outputs": [],
    "source": [
-    "data = pd.read_csv(\"../input/creditcard.csv\")"
+    "data = pd.read_csv(\"./input/creditcard.csv\")"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 44,
    "metadata": {
     "_cell_guid": "5748a40a-3665-4ae6-8b21-36b0dedc2f83",
-    "_uuid": "d800b296c079bf6e11d874cd482058c7fe9db931",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "#data['Normal']=1-data['Class'], instead I am converting Class to categorical\n",
-    "data['Amount_max_fraud'] = 1\n",
-    "data.loc[data.Amount <= 2125.87, 'Amount_max_fraud'] = 0\n",
-    "data.head()"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "fd799d34-4e12-4fff-8b0b-2f1006722544",
-    "_uuid": "dd4cc5fcc249a2d2cc7324ed68e377accdc4fd3e"
-   },
-   "source": [
-    "**Stratify Class column** in train -test split to keep the same Fraud/Normal ratio in train and test data.\n",
-    "I am not using any validation dataset here."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "b2967228-de84-4f0a-b1bb-502d3d45c28d",
-    "_uuid": "87e7f5a83c48a98ae52cc55cf805cdf9a57d8426",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "train,test=train_test_split(data,test_size=0.2,random_state=0,stratify=data['Class'])# stratify the Class"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "0c2923f1-59a3-45b5-ab73-1ffc250ddf40",
-    "_uuid": "0c430eecba7217f93d1ea970249fbbe85cd5f137",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "count_train = pd.value_counts(train['Class'], sort = True).sort_index()\n",
-    "count_test = pd.value_counts(test['Class'], sort = True).sort_index()\n",
-    "print (count_train) \n",
-    "'\\n'  \n",
-    "print(count_test)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "21f6399e-0394-4d33-a7f8-798cd3d6dbd6",
-    "_uuid": "2409d161a9e8233a31e82f81b4cc59d9530c3bc6"
-   },
-   "source": [
-    "Drop target columns from model input datsets"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "78dcafd6-c7e5-4396-b4f4-cba7ae30de9f",
-    "_uuid": "e9eb5647ea9a1daf4f91f4577f3d4d0e8064f096",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "X_train = train.drop(['Class'], axis = 1)\n",
-    "X_test = test.drop(['Class'], axis = 1)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "134879e1-a328-4be0-87ac-a1e5bebc9f6f",
-    "_uuid": "ccc85f3af1d26a61bd3844d0a26110ba341a3572"
-   },
-   "source": [
-    "Define target sets:"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "952ab8e5-cccd-4188-8690-c33e405545d4",
-    "_uuid": "7ae50c4c16bf36e9b001f658909116004a86aed3",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "Y_train = train.loc[:, ['Class']]\n",
-    "Y_test = test.loc[:, ['Class']]"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "5c03b27c-581a-452b-86dd-ae4ce2b37877",
-    "_uuid": "92cce9a63827d1a9e11bc143b46f605f3879eeae",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "# Just sanity check\n",
-    "print(np.shape(X_train))\n",
-    "print(np.shape(Y_train))\n",
-    "print(np.shape(X_test))\n",
-    "print(np.shape(Y_test))"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "51e117b1-cc6f-4855-b0b9-e716f933d3c4",
-    "_uuid": "a784732de39c7687270cfaa734a6f77942132443"
-   },
-   "source": [
-    "Now convert Y_train and Y_test to categorical values with 2 classes."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "94bb1a46-193f-47fd-bf1f-f0e917492298",
-    "_uuid": "add388ac1620f9fb2eb1850b35a81411c30055a6",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "Y_train = to_categorical(Y_train, num_classes = 2)\n",
-    "Y_test = to_categorical(Y_test, num_classes = 2)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "cb55954d-4632-4f06-9d0c-a81e560f9b7f",
-    "_uuid": "078fca56926f75046c3b3e8d4b0e8d19a42b0508"
-   },
-   "source": [
-    "Centering and scaling of input datasets"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "9136cfd3-7e69-4fdf-a658-d0677cb32e26",
-    "_uuid": "94754afc163a673cfc20a64ae01032d0b6e60454",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "#Names of all of the features in X_train.\n",
-    "features = X_train.columns.values\n",
-    "\n",
-    "for feature in features:\n",
-    "    mean, std = data[feature].mean(), data[feature].std()\n",
-    "    X_train.loc[:, feature] = (X_train[feature] - mean) / std\n",
-    "    X_test.loc[:, feature] = (X_test[feature] - mean) / std"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "77bce954-a51b-486c-9f6d-df2510b69c47",
-    "_uuid": "1ee9f2cf77a99b1fbf99cbe70e2ba61595c5bc08",
-    "collapsed": true
-   },
-   "source": [
-    "Now we start Keras model building"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "6c0f7416-412a-4415-94e8-362fa13948ea",
-    "_uuid": "9c648e3b03efea588dea1d20893bf148151447f7",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "# fix random seed for reproducibility\n",
-    "np.random.seed(2)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "92b9a9f3-a529-4ce1-8882-be4c26bac2a5",
-    "_uuid": "4adc5dc848bd35cd6cc78cae6e5a2d21c2ab750d"
-   },
-   "source": [
-    "Set up a 5 layer network with last layer being the output layer.\n",
-    "First layer has input dimention as 31 (number of columns in X_train).\n",
-    "Activation is relu except last layer is with softmax activation\n",
-    "Each layer has dropout at 0.9 (90% of data used at each layer)"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "bbf4cf25-3052-4b8f-9ac7-700865de4905",
-    "_uuid": "cf449dc135ffd673be7ef9b800df733c6c48831d",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "model = Sequential()\n",
-    "model.add(Dense(64, input_dim=31, activation='relu'))\n",
-    "model.add(Dropout(0.9))\n",
-    "model.add(Dense(64, activation='relu'))\n",
-    "model.add(Dropout(0.9))\n",
-    "model.add(Dense(64, activation='relu'))\n",
-    "model.add(Dropout(0.9))\n",
-    "model.add(Dense(64, activation='relu'))\n",
-    "model.add(Dropout(0.9))\n",
-    "model.add(Dense(2, activation='softmax'))  # With 2 outputs"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "f05652b6-31d8-4da9-a9e9-9bd5d9a8df5a",
-    "_uuid": "348f00981f35f87c22ccdac414d86de99f3ac0a1"
-   },
-   "source": [
-    "Compile model using binary crossentropy loss and adam optimizer for loss.\n",
-    "Collect accuracy in metric."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "2dec00ca-e47c-4142-89bc-e7a445314ebc",
-    "_uuid": "086c950d76895f638b313701200a144fdf3851b6",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {
-    "_cell_guid": "88ca1b60-db7e-4994-8c2c-88c788538033",
-    "_uuid": "c708ccee01d5faa3935c01742603e2d9870c968f"
+    "_uuid": "d800b296c079bf6e11d874cd482058c7fe9db931"
    },
-   "source": [
-    "Fit the compiled model on training data.\n",
-    "I am using only 10 epochs to save time with batch_size of 2048."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "metadata": {
-    "_cell_guid": "bbbd50fb-41e8-4f71-b497-f8e19038dad0",
-    "_uuid": "239658b1175b942a7aa179f8ae8b71dccbcb8f6d",
-    "collapsed": true
-   },
-   "outputs": [],
-   "source": [
-    "epoch = 10\n",
-    "batch_size = 2048\n",
-    "model.fit(X_train, Y_train, epochs=epoch, batch_size=batch_size)"
-   ]
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>Time</th>\n",
+       "      <th>V1</th>\n",
+       "      <th>V2</th>\n",
+       "      <th>V3</th>\n",
+       "      <th>V4</th>\n",
+       "      <th>V5</th>\n",
+       "      <th>V6</th>\n",
+       "      <th>V7</th>\n",
+       "      <th>V8</th>\n",
+       "      <th>V9</th>\n",
+       "      <th>V10</th>\n",
+       "      <th>V11</th>\n",
+       "      <th>V12</th>\n",
+       "      <th>V13</th>\n",
+       "      <th>V14</th>\n",
+       "      <th>V15</th>\n",
+       "      <th>V16</th>\n",
+       "      <th>V17</th>\n",
+       "      <th>V18</th>\n",
+       "      <th>V19</th>\n",
+       "      <th>V20</th>\n",
+       "      <th>V21</th>\n",
+       "      <th>V22</th>\n",
+       "      <th>V23</th>\n",
+       "      <th>V24</th>\n",
+       "      <th>V25</th>\n",
+       "      <th>V26</th>\n",
+       "      <th>V27</th>\n",
+       "      <th>V28</th>\n",
+       "      <th>Amount</th>\n",
+       "      <th>Class</th>\n",
+       "      <th>Amount_max_fraud</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>0.0</td>\n",
+       "      <td>-1.359807</td>\n",
+       "      <td>-0.072781</td>\n",
+       "      <td>2.536347</td>\n",
+       "      <td>1.378155</td>\n",
+       "      <td>-0.338321</td>\n",
+       "      <td>0.462388</td>\n",
+       "      <td>0.239599</td>\n",
+       "      <td>0.098698</td>\n",
+       "      <td>0.363787</td>\n",
+       "      <td>0.090794</td>\n",
+       "      <td>-0.551600</td>\n",
+       "      <td>-0.617801</td>\n",
+       "      <td>-0.991390</td>\n",
+       "      <td>-0.311169</td>\n",
+       "      <td>1.468177</td>\n",
+       "      <td>-0.470401</td>\n",
+       "      <td>0.207971</td>\n",
+       "      <td>0.025791</td>\n",
+       "      <td>0.403993</td>\n",
+       "      <td>0.251412</td>\n",
+       "      <td>-0.018307</td>\n",
+       "      <td>0.277838</td>\n",
+       "      <td>-0.110474</td>\n",
+       "      <td>0.066928</td>\n",
+       "      <td>0.128539</td>\n",
+       "      <td>-0.189115</td>\n",
+       "      <td>0.133558</td>\n",
+       "      <td>-0.021053</td>\n",
+       "      <td>149.62</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>0.0</td>\n",
+       "      <td>1.191857</td>\n",
+       "      <td>0.266151</td>\n",
+       "      <td>0.166480</td>\n",
+       "      <td>0.448154</td>\n",
+       "      <td>0.060018</td>\n",
+       "      <td>-0.082361</td>\n",
+       "      <td>-0.078803</td>\n",
+       "      <td>0.085102</td>\n",
+       "      <td>-0.255425</td>\n",
+       "      <td>-0.166974</td>\n",
+       "      <td>1.612727</td>\n",
+       "      <td>1.065235</td>\n",
+       "      <td>0.489095</td>\n",
+       "      <td>-0.143772</td>\n",
+       "      <td>0.635558</td>\n",
+       "      <td>0.463917</td>\n",
+       "      <td>-0.114805</td>\n",
+       "      <td>-0.183361</td>\n",
+       "      <td>-0.145783</td>\n",
+       "      <td>-0.069083</td>\n",
+       "      <td>-0.225775</td>\n",
+       "      <td>-0.638672</td>\n",
+       "      <td>0.101288</td>\n",
+       "      <td>-0.339846</td>\n",
+       "      <td>0.167170</td>\n",
+       "      <td>0.125895</td>\n",
+       "      <td>-0.008983</td>\n",
+       "      <td>0.014724</td>\n",
+       "      <td>2.69</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>1.0</td>\n",
+       "      <td>-1.358354</td>\n",
+       "      <td>-1.340163</td>\n",
+       "      <td>1.773209</td>\n",
+       "      <td>0.379780</td>\n",
+       "      <td>-0.503198</td>\n",
+       "      <td>1.800499</td>\n",
+       "      <td>0.791461</td>\n",
+       "      <td>0.247676</td>\n",
+       "      <td>-1.514654</td>\n",
+       "      <td>0.207643</td>\n",
+       "      <td>0.624501</td>\n",
+       "      <td>0.066084</td>\n",
+       "      <td>0.717293</td>\n",
+       "      <td>-0.165946</td>\n",
+       "      <td>2.345865</td>\n",
+       "      <td>-2.890083</td>\n",
+       "      <td>1.109969</td>\n",
+       "      <td>-0.121359</td>\n",
+       "      <td>-2.261857</td>\n",
+       "      <td>0.524980</td>\n",
+       "      <td>0.247998</td>\n",
+       "      <td>0.771679</td>\n",
+       "      <td>0.909412</td>\n",
+       "      <td>-0.689281</td>\n",
+       "      <td>-0.327642</td>\n",
+       "      <td>-0.139097</td>\n",
+       "      <td>-0.055353</td>\n",
+       "      <td>-0.059752</td>\n",
+       "      <td>378.66</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>1.0</td>\n",
+       "      <td>-0.966272</td>\n",
+       "      <td>-0.185226</td>\n",
+       "      <td>1.792993</td>\n",
+       "      <td>-0.863291</td>\n",
+       "      <td>-0.010309</td>\n",
+       "      <td>1.247203</td>\n",
+       "      <td>0.237609</td>\n",
+       "      <td>0.377436</td>\n",
+       "      <td>-1.387024</td>\n",
+       "      <td>-0.054952</td>\n",
+       "      <td>-0.226487</td>\n",
+       "      <td>0.178228</td>\n",
+       "      <td>0.507757</td>\n",
+       "      <td>-0.287924</td>\n",
+       "      <td>-0.631418</td>\n",
+       "      <td>-1.059647</td>\n",
+       "      <td>-0.684093</td>\n",
+       "      <td>1.965775</td>\n",
+       "      <td>-1.232622</td>\n",
+       "      <td>-0.208038</td>\n",
+       "      <td>-0.108300</td>\n",
+       "      <td>0.005274</td>\n",
+       "      <td>-0.190321</td>\n",
+       "      <td>-1.175575</td>\n",
+       "      <td>0.647376</td>\n",
+       "      <td>-0.221929</td>\n",
+       "      <td>0.062723</td>\n",
+       "      <td>0.061458</td>\n",
+       "      <td>123.50</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>2.0</td>\n",
+       "      <td>-1.158233</td>\n",
+       "      <td>0.877737</td>\n",
+       "      <td>1.548718</td>\n",
+       "      <td>0.403034</td>\n",
+       "      <td>-0.407193</td>\n",
+       "      <td>0.095921</td>\n",
+       "      <td>0.592941</td>\n",
+       "      <td>-0.270533</td>\n",
+       "      <td>0.817739</td>\n",
+       "      <td>0.753074</td>\n",
+       "      <td>-0.822843</td>\n",
+       "      <td>0.538196</td>\n",
+       "      <td>1.345852</td>\n",
+       "      <td>-1.119670</td>\n",
+       "      <td>0.175121</td>\n",
+       "      <td>-0.451449</td>\n",
+       "      <td>-0.237033</td>\n",
+       "      <td>-0.038195</td>\n",
+       "      <td>0.803487</td>\n",
+       "      <td>0.408542</td>\n",
+       "      <td>-0.009431</td>\n",
+       "      <td>0.798278</td>\n",
+       "      <td>-0.137458</td>\n",
+       "      <td>0.141267</td>\n",
+       "      <td>-0.206010</td>\n",
+       "      <td>0.502292</td>\n",
+       "      <td>0.219422</td>\n",
+       "      <td>0.215153</td>\n",
+       "      <td>69.99</td>\n",
+       "      <td>0</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "   Time        V1        V2        V3        V4        V5        V6        V7  \\\n",
+       "0   0.0 -1.359807 -0.072781  2.536347  1.378155 -0.338321  0.462388  0.239599   \n",
+       "1   0.0  1.191857  0.266151  0.166480  0.448154  0.060018 -0.082361 -0.078803   \n",
+       "2   1.0 -1.358354 -1.340163  1.773209  0.379780 -0.503198  1.800499  0.791461   \n",
+       "3   1.0 -0.966272 -0.185226  1.792993 -0.863291 -0.010309  1.247203  0.237609   \n",
+       "4   2.0 -1.158233  0.877737  1.548718  0.403034 -0.407193  0.095921  0.592941   \n",
+       "\n",
+       "         V8        V9       V10       V11       V12       V13       V14  \\\n",
+       "0  0.098698  0.363787  0.090794 -0.551600 -0.617801 -0.991390 -0.311169   \n",
+       "1  0.085102 -0.255425 -0.166974  1.612727  1.065235  0.489095 -0.143772   \n",
+       "2  0.247676 -1.514654  0.207643  0.624501  0.066084  0.717293 -0.165946   \n",
+       "3  0.377436 -1.387024 -0.054952 -0.226487  0.178228  0.507757 -0.287924   \n",
+       "4 -0.270533  0.817739  0.753074 -0.822843  0.538196  1.345852 -1.119670   \n",
+       "\n",
+       "        V15       V16       V17       V18       V19       V20       V21  \\\n",
+       "0  1.468177 -0.470401  0.207971  0.025791  0.403993  0.251412 -0.018307   \n",
+       "1  0.635558  0.463917 -0.114805 -0.183361 -0.145783 -0.069083 -0.225775   \n",
+       "2  2.345865 -2.890083  1.109969 -0.121359 -2.261857  0.524980  0.247998   \n",
+       "3 -0.631418 -1.059647 -0.684093  1.965775 -1.232622 -0.208038 -0.108300   \n",
+       "4  0.175121 -0.451449 -0.237033 -0.038195  0.803487  0.408542 -0.009431   \n",
+       "\n",
+       "        V22       V23       V24       V25       V26       V27       V28  \\\n",
+       "0  0.277838 -0.110474  0.066928  0.128539 -0.189115  0.133558 -0.021053   \n",
+       "1 -0.638672  0.101288 -0.339846  0.167170  0.125895 -0.008983  0.014724   \n",
+       "2  0.771679  0.909412 -0.689281 -0.327642 -0.139097 -0.055353 -0.059752   \n",
+       "3  0.005274 -0.190321 -1.175575  0.647376 -0.221929  0.062723  0.061458   \n",
+       "4  0.798278 -0.137458  0.141267 -0.206010  0.502292  0.219422  0.215153   \n",
+       "\n",
+       "   Amount  Class  Amount_max_fraud  \n",
+       "0  149.62      0                 0  \n",
+       "1    2.69      0                 0  \n",
+       "2  378.66      0                 0  \n",
+       "3  123.50      0                 0  \n",
+       "4   69.99      0                 0  "
+      ]
+     },
+     "execution_count": 44,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "#data['Normal']=1-data['Class'], instead I am converting Class to categorical\n",
+    "data['Amount_max_fraud'] = 1\n",
+    "data.loc[data.Amount <= 2125.87, 'Amount_max_fraud'] = 0\n",
+    "data.head()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Train-Test 데이터를 나눌 때, 사기/보통 거래의 비율을 같게하기 위해 **Class 칼럼을 계층화(Stratify)**합니다. 여기에선 Validation 셋을 쓰지 않겠습니다."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 45,
+   "metadata": {
+    "_cell_guid": "b2967228-de84-4f0a-b1bb-502d3d45c28d",
+    "_uuid": "87e7f5a83c48a98ae52cc55cf805cdf9a57d8426"
+   },
+   "outputs": [],
+   "source": [
+    "train,test=train_test_split(data,test_size=0.2,random_state=0,stratify=data['Class'])# stratify the Class"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 46,
+   "metadata": {
+    "_cell_guid": "0c2923f1-59a3-45b5-ab73-1ffc250ddf40",
+    "_uuid": "0c430eecba7217f93d1ea970249fbbe85cd5f137",
+    "scrolled": false
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "0    227451\n",
+      "1       394\n",
+      "Name: Class, dtype: int64\n",
+      "0    56864\n",
+      "1       98\n",
+      "Name: Class, dtype: int64\n"
+     ]
+    }
+   ],
+   "source": [
+    "count_train = pd.value_counts(train['Class'], sort = True).sort_index()\n",
+    "count_test = pd.value_counts(test['Class'], sort = True).sort_index()\n",
+    "print(count_train) \n",
+    "print(count_test)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 48,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "0.001729245759178389\n",
+      "0.0017204452090867595\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(count_train[1] / (count_train[0] + count_train[1])) # Fraud/Nomal Ratio\n",
+    "print(count_test[1] / (count_test[0] + count_test[1])) # Fraud/Nomal Ratio"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "타겟 칼럼(Class)를 입력 데이터셋에서 삭제(Drop)합니다."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 49,
+   "metadata": {
+    "_cell_guid": "78dcafd6-c7e5-4396-b4f4-cba7ae30de9f",
+    "_uuid": "e9eb5647ea9a1daf4f91f4577f3d4d0e8064f096"
+   },
+   "outputs": [],
+   "source": [
+    "X_train = train.drop(['Class'], axis = 1)\n",
+    "X_test = test.drop(['Class'], axis = 1)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 50,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>Time</th>\n",
+       "      <th>V1</th>\n",
+       "      <th>V2</th>\n",
+       "      <th>V3</th>\n",
+       "      <th>V4</th>\n",
+       "      <th>V5</th>\n",
+       "      <th>V6</th>\n",
+       "      <th>V7</th>\n",
+       "      <th>V8</th>\n",
+       "      <th>V9</th>\n",
+       "      <th>V10</th>\n",
+       "      <th>V11</th>\n",
+       "      <th>V12</th>\n",
+       "      <th>V13</th>\n",
+       "      <th>V14</th>\n",
+       "      <th>V15</th>\n",
+       "      <th>V16</th>\n",
+       "      <th>V17</th>\n",
+       "      <th>V18</th>\n",
+       "      <th>V19</th>\n",
+       "      <th>V20</th>\n",
+       "      <th>V21</th>\n",
+       "      <th>V22</th>\n",
+       "      <th>V23</th>\n",
+       "      <th>V24</th>\n",
+       "      <th>V25</th>\n",
+       "      <th>V26</th>\n",
+       "      <th>V27</th>\n",
+       "      <th>V28</th>\n",
+       "      <th>Amount</th>\n",
+       "      <th>Amount_max_fraud</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>36001</th>\n",
+       "      <td>38355.0</td>\n",
+       "      <td>1.043949</td>\n",
+       "      <td>0.318555</td>\n",
+       "      <td>1.045810</td>\n",
+       "      <td>2.805989</td>\n",
+       "      <td>-0.561113</td>\n",
+       "      <td>-0.367956</td>\n",
+       "      <td>0.032736</td>\n",
+       "      <td>-0.042333</td>\n",
+       "      <td>-0.322674</td>\n",
+       "      <td>0.499167</td>\n",
+       "      <td>-0.572665</td>\n",
+       "      <td>0.346009</td>\n",
+       "      <td>-0.047407</td>\n",
+       "      <td>-0.098964</td>\n",
+       "      <td>-0.663284</td>\n",
+       "      <td>0.181411</td>\n",
+       "      <td>-0.124345</td>\n",
+       "      <td>-0.790453</td>\n",
+       "      <td>-0.720944</td>\n",
+       "      <td>-0.084556</td>\n",
+       "      <td>-0.240105</td>\n",
+       "      <td>-0.680315</td>\n",
+       "      <td>0.085328</td>\n",
+       "      <td>0.684812</td>\n",
+       "      <td>0.318620</td>\n",
+       "      <td>-0.204963</td>\n",
+       "      <td>0.001662</td>\n",
+       "      <td>0.037894</td>\n",
+       "      <td>49.67</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>12844</th>\n",
+       "      <td>22555.0</td>\n",
+       "      <td>-1.665159</td>\n",
+       "      <td>0.808440</td>\n",
+       "      <td>1.805627</td>\n",
+       "      <td>1.903416</td>\n",
+       "      <td>-0.821627</td>\n",
+       "      <td>0.934790</td>\n",
+       "      <td>-0.824802</td>\n",
+       "      <td>0.975890</td>\n",
+       "      <td>1.747469</td>\n",
+       "      <td>-0.658751</td>\n",
+       "      <td>1.281502</td>\n",
+       "      <td>-1.430087</td>\n",
+       "      <td>0.372028</td>\n",
+       "      <td>1.403024</td>\n",
+       "      <td>-2.739413</td>\n",
+       "      <td>-1.331766</td>\n",
+       "      <td>1.964590</td>\n",
+       "      <td>-0.205639</td>\n",
+       "      <td>1.325588</td>\n",
+       "      <td>-0.373759</td>\n",
+       "      <td>-0.335332</td>\n",
+       "      <td>-0.510994</td>\n",
+       "      <td>0.035839</td>\n",
+       "      <td>0.147565</td>\n",
+       "      <td>-0.529358</td>\n",
+       "      <td>-0.566950</td>\n",
+       "      <td>-0.595998</td>\n",
+       "      <td>-0.220086</td>\n",
+       "      <td>16.94</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2873</th>\n",
+       "      <td>2431.0</td>\n",
+       "      <td>-0.324096</td>\n",
+       "      <td>0.601836</td>\n",
+       "      <td>0.865329</td>\n",
+       "      <td>-2.138000</td>\n",
+       "      <td>0.294663</td>\n",
+       "      <td>-1.251553</td>\n",
+       "      <td>1.072114</td>\n",
+       "      <td>-0.334896</td>\n",
+       "      <td>1.071268</td>\n",
+       "      <td>-1.109522</td>\n",
+       "      <td>-1.016020</td>\n",
+       "      <td>-0.654945</td>\n",
+       "      <td>-1.473470</td>\n",
+       "      <td>0.317345</td>\n",
+       "      <td>1.067491</td>\n",
+       "      <td>-0.372642</td>\n",
+       "      <td>-0.674725</td>\n",
+       "      <td>0.369841</td>\n",
+       "      <td>0.095583</td>\n",
+       "      <td>-0.039868</td>\n",
+       "      <td>0.012220</td>\n",
+       "      <td>0.352856</td>\n",
+       "      <td>-0.341505</td>\n",
+       "      <td>-0.145791</td>\n",
+       "      <td>0.094194</td>\n",
+       "      <td>-0.804026</td>\n",
+       "      <td>0.229428</td>\n",
+       "      <td>-0.021623</td>\n",
+       "      <td>1.00</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>145263</th>\n",
+       "      <td>86773.0</td>\n",
+       "      <td>-0.258270</td>\n",
+       "      <td>1.217501</td>\n",
+       "      <td>-0.585348</td>\n",
+       "      <td>-0.875347</td>\n",
+       "      <td>1.222481</td>\n",
+       "      <td>-0.311027</td>\n",
+       "      <td>1.073860</td>\n",
+       "      <td>-0.161408</td>\n",
+       "      <td>0.200665</td>\n",
+       "      <td>0.154307</td>\n",
+       "      <td>0.882673</td>\n",
+       "      <td>0.547890</td>\n",
+       "      <td>0.269484</td>\n",
+       "      <td>-1.253302</td>\n",
+       "      <td>-0.883963</td>\n",
+       "      <td>0.495221</td>\n",
+       "      <td>-0.153212</td>\n",
+       "      <td>0.296710</td>\n",
+       "      <td>0.136148</td>\n",
+       "      <td>0.382305</td>\n",
+       "      <td>-0.424626</td>\n",
+       "      <td>-0.781158</td>\n",
+       "      <td>0.019316</td>\n",
+       "      <td>0.178614</td>\n",
+       "      <td>-0.315616</td>\n",
+       "      <td>0.096665</td>\n",
+       "      <td>0.269740</td>\n",
+       "      <td>-0.020635</td>\n",
+       "      <td>10.78</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>186658</th>\n",
+       "      <td>127202.0</td>\n",
+       "      <td>2.142162</td>\n",
+       "      <td>-0.494988</td>\n",
+       "      <td>-1.936511</td>\n",
+       "      <td>-0.818288</td>\n",
+       "      <td>-0.025213</td>\n",
+       "      <td>-1.027245</td>\n",
+       "      <td>-0.151627</td>\n",
+       "      <td>-0.305750</td>\n",
+       "      <td>-0.869482</td>\n",
+       "      <td>0.428729</td>\n",
+       "      <td>1.136666</td>\n",
+       "      <td>0.273476</td>\n",
+       "      <td>0.697123</td>\n",
+       "      <td>-1.222134</td>\n",
+       "      <td>-0.938820</td>\n",
+       "      <td>1.298149</td>\n",
+       "      <td>0.912921</td>\n",
+       "      <td>-0.793721</td>\n",
+       "      <td>1.064984</td>\n",
+       "      <td>0.106592</td>\n",
+       "      <td>0.010115</td>\n",
+       "      <td>0.021722</td>\n",
+       "      <td>0.079463</td>\n",
+       "      <td>-0.480899</td>\n",
+       "      <td>0.023846</td>\n",
+       "      <td>-0.279076</td>\n",
+       "      <td>-0.030121</td>\n",
+       "      <td>-0.043888</td>\n",
+       "      <td>39.96</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>34271</th>\n",
+       "      <td>37590.0</td>\n",
+       "      <td>-6.341667</td>\n",
+       "      <td>1.192815</td>\n",
+       "      <td>-3.686579</td>\n",
+       "      <td>-2.754006</td>\n",
+       "      <td>-5.007164</td>\n",
+       "      <td>3.988118</td>\n",
+       "      <td>1.693606</td>\n",
+       "      <td>-13.577522</td>\n",
+       "      <td>2.557965</td>\n",
+       "      <td>-0.931438</td>\n",
+       "      <td>-1.189333</td>\n",
+       "      <td>0.762725</td>\n",
+       "      <td>-0.517857</td>\n",
+       "      <td>0.067279</td>\n",
+       "      <td>-0.099498</td>\n",
+       "      <td>0.837768</td>\n",
+       "      <td>-0.156679</td>\n",
+       "      <td>-0.402600</td>\n",
+       "      <td>-0.238742</td>\n",
+       "      <td>-7.171342</td>\n",
+       "      <td>9.779321</td>\n",
+       "      <td>-3.660198</td>\n",
+       "      <td>-1.507217</td>\n",
+       "      <td>-0.179185</td>\n",
+       "      <td>0.002092</td>\n",
+       "      <td>-0.019168</td>\n",
+       "      <td>3.611243</td>\n",
+       "      <td>-0.894494</td>\n",
+       "      <td>1437.08</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>76782</th>\n",
+       "      <td>56739.0</td>\n",
+       "      <td>-0.463928</td>\n",
+       "      <td>0.204745</td>\n",
+       "      <td>2.795734</td>\n",
+       "      <td>1.546200</td>\n",
+       "      <td>-0.720374</td>\n",
+       "      <td>1.108212</td>\n",
+       "      <td>-0.581133</td>\n",
+       "      <td>0.544075</td>\n",
+       "      <td>1.099843</td>\n",
+       "      <td>-0.653523</td>\n",
+       "      <td>-1.420838</td>\n",
+       "      <td>0.427211</td>\n",
+       "      <td>-0.436201</td>\n",
+       "      <td>-1.055012</td>\n",
+       "      <td>-1.225903</td>\n",
+       "      <td>-1.179268</td>\n",
+       "      <td>0.846734</td>\n",
+       "      <td>-0.374502</td>\n",
+       "      <td>0.623190</td>\n",
+       "      <td>-0.052393</td>\n",
+       "      <td>0.094431</td>\n",
+       "      <td>0.788161</td>\n",
+       "      <td>-0.100344</td>\n",
+       "      <td>0.126248</td>\n",
+       "      <td>-0.461584</td>\n",
+       "      <td>-0.200985</td>\n",
+       "      <td>0.280168</td>\n",
+       "      <td>0.180046</td>\n",
+       "      <td>11.10</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1169</th>\n",
+       "      <td>912.0</td>\n",
+       "      <td>-1.394619</td>\n",
+       "      <td>0.554224</td>\n",
+       "      <td>-0.617805</td>\n",
+       "      <td>-0.313706</td>\n",
+       "      <td>2.157413</td>\n",
+       "      <td>3.537534</td>\n",
+       "      <td>-0.567305</td>\n",
+       "      <td>1.380888</td>\n",
+       "      <td>-0.112291</td>\n",
+       "      <td>0.107269</td>\n",
+       "      <td>-0.637728</td>\n",
+       "      <td>0.099510</td>\n",
+       "      <td>-0.062533</td>\n",
+       "      <td>0.254824</td>\n",
+       "      <td>0.318805</td>\n",
+       "      <td>-0.335245</td>\n",
+       "      <td>0.052870</td>\n",
+       "      <td>0.048783</td>\n",
+       "      <td>1.654699</td>\n",
+       "      <td>0.336595</td>\n",
+       "      <td>-0.196300</td>\n",
+       "      <td>-0.546172</td>\n",
+       "      <td>-0.124758</td>\n",
+       "      <td>1.046780</td>\n",
+       "      <td>-0.089474</td>\n",
+       "      <td>0.363674</td>\n",
+       "      <td>0.215692</td>\n",
+       "      <td>0.225698</td>\n",
+       "      <td>24.08</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>282674</th>\n",
+       "      <td>171058.0</td>\n",
+       "      <td>-2.121785</td>\n",
+       "      <td>-1.518485</td>\n",
+       "      <td>2.667942</td>\n",
+       "      <td>-2.458839</td>\n",
+       "      <td>-1.414952</td>\n",
+       "      <td>0.807935</td>\n",
+       "      <td>-1.870160</td>\n",
+       "      <td>0.516477</td>\n",
+       "      <td>-0.553359</td>\n",
+       "      <td>0.645799</td>\n",
+       "      <td>-0.514941</td>\n",
+       "      <td>0.618199</td>\n",
+       "      <td>1.461221</td>\n",
+       "      <td>-2.116306</td>\n",
+       "      <td>-3.244732</td>\n",
+       "      <td>-0.345286</td>\n",
+       "      <td>0.368870</td>\n",
+       "      <td>0.699637</td>\n",
+       "      <td>0.031145</td>\n",
+       "      <td>-0.684797</td>\n",
+       "      <td>-0.079591</td>\n",
+       "      <td>0.659885</td>\n",
+       "      <td>-1.000976</td>\n",
+       "      <td>-0.212991</td>\n",
+       "      <td>0.420722</td>\n",
+       "      <td>-0.019645</td>\n",
+       "      <td>-0.482792</td>\n",
+       "      <td>-0.407780</td>\n",
+       "      <td>28.00</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>130872</th>\n",
+       "      <td>79450.0</td>\n",
+       "      <td>-2.254788</td>\n",
+       "      <td>-0.955967</td>\n",
+       "      <td>0.590596</td>\n",
+       "      <td>0.114744</td>\n",
+       "      <td>-5.876531</td>\n",
+       "      <td>3.364858</td>\n",
+       "      <td>6.365904</td>\n",
+       "      <td>-1.252960</td>\n",
+       "      <td>0.243849</td>\n",
+       "      <td>-1.637022</td>\n",
+       "      <td>-0.626434</td>\n",
+       "      <td>0.105457</td>\n",
+       "      <td>0.859541</td>\n",
+       "      <td>-1.306715</td>\n",
+       "      <td>-0.724693</td>\n",
+       "      <td>0.553933</td>\n",
+       "      <td>-0.590110</td>\n",
+       "      <td>-0.471491</td>\n",
+       "      <td>-0.291435</td>\n",
+       "      <td>-0.772127</td>\n",
+       "      <td>-0.601908</td>\n",
+       "      <td>-0.092918</td>\n",
+       "      <td>0.517855</td>\n",
+       "      <td>0.534361</td>\n",
+       "      <td>0.494493</td>\n",
+       "      <td>0.301414</td>\n",
+       "      <td>0.507071</td>\n",
+       "      <td>-0.560888</td>\n",
+       "      <td>1354.98</td>\n",
+       "      <td>0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "            Time        V1        V2        V3        V4        V5        V6  \\\n",
+       "36001    38355.0  1.043949  0.318555  1.045810  2.805989 -0.561113 -0.367956   \n",
+       "12844    22555.0 -1.665159  0.808440  1.805627  1.903416 -0.821627  0.934790   \n",
+       "2873      2431.0 -0.324096  0.601836  0.865329 -2.138000  0.294663 -1.251553   \n",
+       "145263   86773.0 -0.258270  1.217501 -0.585348 -0.875347  1.222481 -0.311027   \n",
+       "186658  127202.0  2.142162 -0.494988 -1.936511 -0.818288 -0.025213 -1.027245   \n",
+       "34271    37590.0 -6.341667  1.192815 -3.686579 -2.754006 -5.007164  3.988118   \n",
+       "76782    56739.0 -0.463928  0.204745  2.795734  1.546200 -0.720374  1.108212   \n",
+       "1169       912.0 -1.394619  0.554224 -0.617805 -0.313706  2.157413  3.537534   \n",
+       "282674  171058.0 -2.121785 -1.518485  2.667942 -2.458839 -1.414952  0.807935   \n",
+       "130872   79450.0 -2.254788 -0.955967  0.590596  0.114744 -5.876531  3.364858   \n",
+       "\n",
+       "              V7         V8        V9       V10       V11       V12       V13  \\\n",
+       "36001   0.032736  -0.042333 -0.322674  0.499167 -0.572665  0.346009 -0.047407   \n",
+       "12844  -0.824802   0.975890  1.747469 -0.658751  1.281502 -1.430087  0.372028   \n",
+       "2873    1.072114  -0.334896  1.071268 -1.109522 -1.016020 -0.654945 -1.473470   \n",
+       "145263  1.073860  -0.161408  0.200665  0.154307  0.882673  0.547890  0.269484   \n",
+       "186658 -0.151627  -0.305750 -0.869482  0.428729  1.136666  0.273476  0.697123   \n",
+       "34271   1.693606 -13.577522  2.557965 -0.931438 -1.189333  0.762725 -0.517857   \n",
+       "76782  -0.581133   0.544075  1.099843 -0.653523 -1.420838  0.427211 -0.436201   \n",
+       "1169   -0.567305   1.380888 -0.112291  0.107269 -0.637728  0.099510 -0.062533   \n",
+       "282674 -1.870160   0.516477 -0.553359  0.645799 -0.514941  0.618199  1.461221   \n",
+       "130872  6.365904  -1.252960  0.243849 -1.637022 -0.626434  0.105457  0.859541   \n",
+       "\n",
+       "             V14       V15       V16       V17       V18       V19       V20  \\\n",
+       "36001  -0.098964 -0.663284  0.181411 -0.124345 -0.790453 -0.720944 -0.084556   \n",
+       "12844   1.403024 -2.739413 -1.331766  1.964590 -0.205639  1.325588 -0.373759   \n",
+       "2873    0.317345  1.067491 -0.372642 -0.674725  0.369841  0.095583 -0.039868   \n",
+       "145263 -1.253302 -0.883963  0.495221 -0.153212  0.296710  0.136148  0.382305   \n",
+       "186658 -1.222134 -0.938820  1.298149  0.912921 -0.793721  1.064984  0.106592   \n",
+       "34271   0.067279 -0.099498  0.837768 -0.156679 -0.402600 -0.238742 -7.171342   \n",
+       "76782  -1.055012 -1.225903 -1.179268  0.846734 -0.374502  0.623190 -0.052393   \n",
+       "1169    0.254824  0.318805 -0.335245  0.052870  0.048783  1.654699  0.336595   \n",
+       "282674 -2.116306 -3.244732 -0.345286  0.368870  0.699637  0.031145 -0.684797   \n",
+       "130872 -1.306715 -0.724693  0.553933 -0.590110 -0.471491 -0.291435 -0.772127   \n",
+       "\n",
+       "             V21       V22       V23       V24       V25       V26       V27  \\\n",
+       "36001  -0.240105 -0.680315  0.085328  0.684812  0.318620 -0.204963  0.001662   \n",
+       "12844  -0.335332 -0.510994  0.035839  0.147565 -0.529358 -0.566950 -0.595998   \n",
+       "2873    0.012220  0.352856 -0.341505 -0.145791  0.094194 -0.804026  0.229428   \n",
+       "145263 -0.424626 -0.781158  0.019316  0.178614 -0.315616  0.096665  0.269740   \n",
+       "186658  0.010115  0.021722  0.079463 -0.480899  0.023846 -0.279076 -0.030121   \n",
+       "34271   9.779321 -3.660198 -1.507217 -0.179185  0.002092 -0.019168  3.611243   \n",
+       "76782   0.094431  0.788161 -0.100344  0.126248 -0.461584 -0.200985  0.280168   \n",
+       "1169   -0.196300 -0.546172 -0.124758  1.046780 -0.089474  0.363674  0.215692   \n",
+       "282674 -0.079591  0.659885 -1.000976 -0.212991  0.420722 -0.019645 -0.482792   \n",
+       "130872 -0.601908 -0.092918  0.517855  0.534361  0.494493  0.301414  0.507071   \n",
+       "\n",
+       "             V28   Amount  Amount_max_fraud  \n",
+       "36001   0.037894    49.67                 0  \n",
+       "12844  -0.220086    16.94                 0  \n",
+       "2873   -0.021623     1.00                 0  \n",
+       "145263 -0.020635    10.78                 0  \n",
+       "186658 -0.043888    39.96                 0  \n",
+       "34271  -0.894494  1437.08                 0  \n",
+       "76782   0.180046    11.10                 0  \n",
+       "1169    0.225698    24.08                 0  \n",
+       "282674 -0.407780    28.00                 0  \n",
+       "130872 -0.560888  1354.98                 0  "
+      ]
+     },
+     "execution_count": 50,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "X_train.head(n=10)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "그리고 타겟을 정의합니다."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 51,
+   "metadata": {
+    "_cell_guid": "952ab8e5-cccd-4188-8690-c33e405545d4",
+    "_uuid": "7ae50c4c16bf36e9b001f658909116004a86aed3"
+   },
+   "outputs": [],
+   "source": [
+    "Y_train = train.loc[:, ['Class']]\n",
+    "Y_test = test.loc[:, ['Class']]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 52,
+   "metadata": {
+    "_cell_guid": "5c03b27c-581a-452b-86dd-ae4ce2b37877",
+    "_uuid": "92cce9a63827d1a9e11bc143b46f605f3879eeae"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "(227845, 31)\n",
+      "(227845, 1)\n",
+      "(56962, 31)\n",
+      "(56962, 1)\n"
+     ]
+    }
+   ],
+   "source": [
+    "# Just sanity check\n",
+    "print(np.shape(X_train))\n",
+    "print(np.shape(Y_train))\n",
+    "print(np.shape(X_test))\n",
+    "print(np.shape(Y_test))"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "이번엔 Y_train과 Y_Test를 2개의 클래스를 가지는 Categorical Values로 변환합니다."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 53,
+   "metadata": {
+    "_cell_guid": "94bb1a46-193f-47fd-bf1f-f0e917492298",
+    "_uuid": "add388ac1620f9fb2eb1850b35a81411c30055a6"
+   },
+   "outputs": [],
+   "source": [
+    "Y_train = to_categorical(Y_train, num_classes = 2)\n",
+    "Y_test = to_categorical(Y_test, num_classes = 2)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 65,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "array([[0., 1.],\n",
+       "       [1., 0.]], dtype=float32)"
+      ]
+     },
+     "execution_count": 65,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "np.unique(Y_train, axis=0)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "입력 데이터셋에 정규화(Centering and Scaling)를 적용합니다."
+   ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 59,
    "metadata": {
-    "_cell_guid": "e45c27c0-1e29-47cf-befe-bba155dd21cf",
-    "_uuid": "a6790157aaf0370faa31223ebd4c1faeeaf84f9a",
-    "collapsed": true
+    "_cell_guid": "9136cfd3-7e69-4fdf-a658-d0677cb32e26",
+    "_uuid": "94754afc163a673cfc20a64ae01032d0b6e60454"
+   },
+   "outputs": [],
+   "source": [
+    "#Names of all of the features in X_train.\n",
+    "features = X_train.columns.values\n",
+    "\n",
+    "for feature in features:\n",
+    "    mean, std = data[feature].mean(), data[feature].std()\n",
+    "    X_train.loc[:, feature] = (X_train[feature] - mean) / std\n",
+    "    X_test.loc[:, feature] = (X_test[feature] - mean) / std"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 60,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>Time</th>\n",
+       "      <th>V1</th>\n",
+       "      <th>V2</th>\n",
+       "      <th>V3</th>\n",
+       "      <th>V4</th>\n",
+       "      <th>V5</th>\n",
+       "      <th>V6</th>\n",
+       "      <th>V7</th>\n",
+       "      <th>V8</th>\n",
+       "      <th>V9</th>\n",
+       "      <th>V10</th>\n",
+       "      <th>V11</th>\n",
+       "      <th>V12</th>\n",
+       "      <th>V13</th>\n",
+       "      <th>V14</th>\n",
+       "      <th>V15</th>\n",
+       "      <th>V16</th>\n",
+       "      <th>V17</th>\n",
+       "      <th>V18</th>\n",
+       "      <th>V19</th>\n",
+       "      <th>V20</th>\n",
+       "      <th>V21</th>\n",
+       "      <th>V22</th>\n",
+       "      <th>V23</th>\n",
+       "      <th>V24</th>\n",
+       "      <th>V25</th>\n",
+       "      <th>V26</th>\n",
+       "      <th>V27</th>\n",
+       "      <th>V28</th>\n",
+       "      <th>Amount</th>\n",
+       "      <th>Amount_max_fraud</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>count</th>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "      <td>56962.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>mean</th>\n",
+       "      <td>0.005139</td>\n",
+       "      <td>0.006783</td>\n",
+       "      <td>0.004003</td>\n",
+       "      <td>-0.002811</td>\n",
+       "      <td>0.001057</td>\n",
+       "      <td>-0.002541</td>\n",
+       "      <td>-0.002311</td>\n",
+       "      <td>0.000114</td>\n",
+       "      <td>-0.005443</td>\n",
+       "      <td>0.001423</td>\n",
+       "      <td>0.002917</td>\n",
+       "      <td>-0.008161</td>\n",
+       "      <td>-0.000041</td>\n",
+       "      <td>-0.000321</td>\n",
+       "      <td>-0.003872</td>\n",
+       "      <td>0.003222</td>\n",
+       "      <td>-0.001978</td>\n",
+       "      <td>0.000035</td>\n",
+       "      <td>0.003967</td>\n",
+       "      <td>0.000940</td>\n",
+       "      <td>-0.003482</td>\n",
+       "      <td>-0.003067</td>\n",
+       "      <td>-0.006804</td>\n",
+       "      <td>0.006418</td>\n",
+       "      <td>-0.001680</td>\n",
+       "      <td>-0.001673</td>\n",
+       "      <td>0.009358</td>\n",
+       "      <td>0.003430</td>\n",
+       "      <td>-0.006029</td>\n",
+       "      <td>-0.002762</td>\n",
+       "      <td>-0.005195</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>std</th>\n",
+       "      <td>0.998959</td>\n",
+       "      <td>0.991084</td>\n",
+       "      <td>0.975726</td>\n",
+       "      <td>1.000395</td>\n",
+       "      <td>1.002287</td>\n",
+       "      <td>1.037301</td>\n",
+       "      <td>1.020548</td>\n",
+       "      <td>1.052306</td>\n",
+       "      <td>1.005681</td>\n",
+       "      <td>1.004602</td>\n",
+       "      <td>1.005792</td>\n",
+       "      <td>0.995285</td>\n",
+       "      <td>0.998488</td>\n",
+       "      <td>0.999305</td>\n",
+       "      <td>0.995902</td>\n",
+       "      <td>0.996960</td>\n",
+       "      <td>1.002095</td>\n",
+       "      <td>0.991084</td>\n",
+       "      <td>0.997565</td>\n",
+       "      <td>1.006987</td>\n",
+       "      <td>0.991601</td>\n",
+       "      <td>1.001834</td>\n",
+       "      <td>1.006337</td>\n",
+       "      <td>0.995456</td>\n",
+       "      <td>0.997595</td>\n",
+       "      <td>0.999471</td>\n",
+       "      <td>0.999232</td>\n",
+       "      <td>1.032498</td>\n",
+       "      <td>0.986538</td>\n",
+       "      <td>1.031676</td>\n",
+       "      <td>0.942043</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>min</th>\n",
+       "      <td>-1.996537</td>\n",
+       "      <td>-18.640252</td>\n",
+       "      <td>-28.722479</td>\n",
+       "      <td>-31.871677</td>\n",
+       "      <td>-3.927001</td>\n",
+       "      <td>-82.407952</td>\n",
+       "      <td>-15.052954</td>\n",
+       "      <td>-22.807581</td>\n",
+       "      <td>-34.734141</td>\n",
+       "      <td>-8.630238</td>\n",
+       "      <td>-19.239745</td>\n",
+       "      <td>-3.927948</td>\n",
+       "      <td>-18.568526</td>\n",
+       "      <td>-3.867219</td>\n",
+       "      <td>-20.044245</td>\n",
+       "      <td>-4.470963</td>\n",
+       "      <td>-14.449819</td>\n",
+       "      <td>-25.561333</td>\n",
+       "      <td>-10.569916</td>\n",
+       "      <td>-5.674207</td>\n",
+       "      <td>-70.691337</td>\n",
+       "      <td>-29.434190</td>\n",
+       "      <td>-11.841840</td>\n",
+       "      <td>-52.571789</td>\n",
+       "      <td>-4.660608</td>\n",
+       "      <td>-16.683278</td>\n",
+       "      <td>-3.687188</td>\n",
+       "      <td>-21.996904</td>\n",
+       "      <td>-46.746035</td>\n",
+       "      <td>-0.353229</td>\n",
+       "      <td>-0.046062</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>25%</th>\n",
+       "      <td>-0.853426</td>\n",
+       "      <td>-0.463875</td>\n",
+       "      <td>-0.363474</td>\n",
+       "      <td>-0.589585</td>\n",
+       "      <td>-0.598655</td>\n",
+       "      <td>-0.502851</td>\n",
+       "      <td>-0.578069</td>\n",
+       "      <td>-0.445477</td>\n",
+       "      <td>-0.176824</td>\n",
+       "      <td>-0.584429</td>\n",
+       "      <td>-0.491456</td>\n",
+       "      <td>-0.752641</td>\n",
+       "      <td>-0.406027</td>\n",
+       "      <td>-0.651183</td>\n",
+       "      <td>-0.442192</td>\n",
+       "      <td>-0.629025</td>\n",
+       "      <td>-0.538465</td>\n",
+       "      <td>-0.570989</td>\n",
+       "      <td>-0.590404</td>\n",
+       "      <td>-0.559835</td>\n",
+       "      <td>-0.274473</td>\n",
+       "      <td>-0.313414</td>\n",
+       "      <td>-0.756551</td>\n",
+       "      <td>-0.256508</td>\n",
+       "      <td>-0.588789</td>\n",
+       "      <td>-0.605699</td>\n",
+       "      <td>-0.669323</td>\n",
+       "      <td>-0.173930</td>\n",
+       "      <td>-0.158599</td>\n",
+       "      <td>-0.330780</td>\n",
+       "      <td>-0.046062</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>50%</th>\n",
+       "      <td>-0.201890</td>\n",
+       "      <td>0.018508</td>\n",
+       "      <td>0.037449</td>\n",
+       "      <td>0.112718</td>\n",
+       "      <td>-0.015604</td>\n",
+       "      <td>-0.037550</td>\n",
+       "      <td>-0.209371</td>\n",
+       "      <td>0.031651</td>\n",
+       "      <td>0.016664</td>\n",
+       "      <td>-0.047884</td>\n",
+       "      <td>-0.087869</td>\n",
+       "      <td>-0.042653</td>\n",
+       "      <td>0.140433</td>\n",
+       "      <td>-0.013747</td>\n",
+       "      <td>0.048731</td>\n",
+       "      <td>0.056303</td>\n",
+       "      <td>0.075434</td>\n",
+       "      <td>-0.078295</td>\n",
+       "      <td>-0.006511</td>\n",
+       "      <td>0.004373</td>\n",
+       "      <td>-0.080084</td>\n",
+       "      <td>-0.044081</td>\n",
+       "      <td>-0.000454</td>\n",
+       "      <td>-0.014947</td>\n",
+       "      <td>0.066119</td>\n",
+       "      <td>0.034672</td>\n",
+       "      <td>-0.099246</td>\n",
+       "      <td>0.003594</td>\n",
+       "      <td>0.034397</td>\n",
+       "      <td>-0.265671</td>\n",
+       "      <td>-0.046062</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>75%</th>\n",
+       "      <td>0.935415</td>\n",
+       "      <td>0.673778</td>\n",
+       "      <td>0.485168</td>\n",
+       "      <td>0.670664</td>\n",
+       "      <td>0.521721</td>\n",
+       "      <td>0.445901</td>\n",
+       "      <td>0.294835</td>\n",
+       "      <td>0.460571</td>\n",
+       "      <td>0.272035</td>\n",
+       "      <td>0.545800</td>\n",
+       "      <td>0.416090</td>\n",
+       "      <td>0.713728</td>\n",
+       "      <td>0.617349</td>\n",
+       "      <td>0.671185</td>\n",
+       "      <td>0.508547</td>\n",
+       "      <td>0.708872</td>\n",
+       "      <td>0.594398</td>\n",
+       "      <td>0.474034</td>\n",
+       "      <td>0.599056</td>\n",
+       "      <td>0.563700</td>\n",
+       "      <td>0.170239</td>\n",
+       "      <td>0.251604</td>\n",
+       "      <td>0.722727</td>\n",
+       "      <td>0.240184</td>\n",
+       "      <td>0.722627</td>\n",
+       "      <td>0.669307</td>\n",
+       "      <td>0.508652</td>\n",
+       "      <td>0.224801</td>\n",
+       "      <td>0.234739</td>\n",
+       "      <td>-0.043378</td>\n",
+       "      <td>-0.046062</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>max</th>\n",
+       "      <td>1.641907</td>\n",
+       "      <td>1.249048</td>\n",
+       "      <td>11.446954</td>\n",
+       "      <td>2.628480</td>\n",
+       "      <td>10.809043</td>\n",
+       "      <td>21.022418</td>\n",
+       "      <td>55.020053</td>\n",
+       "      <td>97.478068</td>\n",
+       "      <td>15.307174</td>\n",
+       "      <td>9.439610</td>\n",
+       "      <td>13.992774</td>\n",
+       "      <td>11.775017</td>\n",
+       "      <td>4.850326</td>\n",
+       "      <td>7.160723</td>\n",
+       "      <td>7.742518</td>\n",
+       "      <td>9.699100</td>\n",
+       "      <td>19.760405</td>\n",
+       "      <td>9.105285</td>\n",
+       "      <td>4.883746</td>\n",
+       "      <td>6.869402</td>\n",
+       "      <td>25.613973</td>\n",
+       "      <td>30.788496</td>\n",
+       "      <td>9.357086</td>\n",
+       "      <td>33.314118</td>\n",
+       "      <td>7.569671</td>\n",
+       "      <td>14.425293</td>\n",
+       "      <td>7.293962</td>\n",
+       "      <td>78.319259</td>\n",
+       "      <td>68.528383</td>\n",
+       "      <td>102.362063</td>\n",
+       "      <td>21.709793</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "               Time            V1            V2            V3            V4  \\\n",
+       "count  56962.000000  56962.000000  56962.000000  56962.000000  56962.000000   \n",
+       "mean       0.005139      0.006783      0.004003     -0.002811      0.001057   \n",
+       "std        0.998959      0.991084      0.975726      1.000395      1.002287   \n",
+       "min       -1.996537    -18.640252    -28.722479    -31.871677     -3.927001   \n",
+       "25%       -0.853426     -0.463875     -0.363474     -0.589585     -0.598655   \n",
+       "50%       -0.201890      0.018508      0.037449      0.112718     -0.015604   \n",
+       "75%        0.935415      0.673778      0.485168      0.670664      0.521721   \n",
+       "max        1.641907      1.249048     11.446954      2.628480     10.809043   \n",
+       "\n",
+       "                 V5            V6            V7            V8            V9  \\\n",
+       "count  56962.000000  56962.000000  56962.000000  56962.000000  56962.000000   \n",
+       "mean      -0.002541     -0.002311      0.000114     -0.005443      0.001423   \n",
+       "std        1.037301      1.020548      1.052306      1.005681      1.004602   \n",
+       "min      -82.407952    -15.052954    -22.807581    -34.734141     -8.630238   \n",
+       "25%       -0.502851     -0.578069     -0.445477     -0.176824     -0.584429   \n",
+       "50%       -0.037550     -0.209371      0.031651      0.016664     -0.047884   \n",
+       "75%        0.445901      0.294835      0.460571      0.272035      0.545800   \n",
+       "max       21.022418     55.020053     97.478068     15.307174      9.439610   \n",
+       "\n",
+       "                V10           V11           V12           V13           V14  \\\n",
+       "count  56962.000000  56962.000000  56962.000000  56962.000000  56962.000000   \n",
+       "mean       0.002917     -0.008161     -0.000041     -0.000321     -0.003872   \n",
+       "std        1.005792      0.995285      0.998488      0.999305      0.995902   \n",
+       "min      -19.239745     -3.927948    -18.568526     -3.867219    -20.044245   \n",
+       "25%       -0.491456     -0.752641     -0.406027     -0.651183     -0.442192   \n",
+       "50%       -0.087869     -0.042653      0.140433     -0.013747      0.048731   \n",
+       "75%        0.416090      0.713728      0.617349      0.671185      0.508547   \n",
+       "max       13.992774     11.775017      4.850326      7.160723      7.742518   \n",
+       "\n",
+       "                V15           V16           V17           V18           V19  \\\n",
+       "count  56962.000000  56962.000000  56962.000000  56962.000000  56962.000000   \n",
+       "mean       0.003222     -0.001978      0.000035      0.003967      0.000940   \n",
+       "std        0.996960      1.002095      0.991084      0.997565      1.006987   \n",
+       "min       -4.470963    -14.449819    -25.561333    -10.569916     -5.674207   \n",
+       "25%       -0.629025     -0.538465     -0.570989     -0.590404     -0.559835   \n",
+       "50%        0.056303      0.075434     -0.078295     -0.006511      0.004373   \n",
+       "75%        0.708872      0.594398      0.474034      0.599056      0.563700   \n",
+       "max        9.699100     19.760405      9.105285      4.883746      6.869402   \n",
+       "\n",
+       "                V20           V21           V22           V23           V24  \\\n",
+       "count  56962.000000  56962.000000  56962.000000  56962.000000  56962.000000   \n",
+       "mean      -0.003482     -0.003067     -0.006804      0.006418     -0.001680   \n",
+       "std        0.991601      1.001834      1.006337      0.995456      0.997595   \n",
+       "min      -70.691337    -29.434190    -11.841840    -52.571789     -4.660608   \n",
+       "25%       -0.274473     -0.313414     -0.756551     -0.256508     -0.588789   \n",
+       "50%       -0.080084     -0.044081     -0.000454     -0.014947      0.066119   \n",
+       "75%        0.170239      0.251604      0.722727      0.240184      0.722627   \n",
+       "max       25.613973     30.788496      9.357086     33.314118      7.569671   \n",
+       "\n",
+       "                V25           V26           V27           V28        Amount  \\\n",
+       "count  56962.000000  56962.000000  56962.000000  56962.000000  56962.000000   \n",
+       "mean      -0.001673      0.009358      0.003430     -0.006029     -0.002762   \n",
+       "std        0.999471      0.999232      1.032498      0.986538      1.031676   \n",
+       "min      -16.683278     -3.687188    -21.996904    -46.746035     -0.353229   \n",
+       "25%       -0.605699     -0.669323     -0.173930     -0.158599     -0.330780   \n",
+       "50%        0.034672     -0.099246      0.003594      0.034397     -0.265671   \n",
+       "75%        0.669307      0.508652      0.224801      0.234739     -0.043378   \n",
+       "max       14.425293      7.293962     78.319259     68.528383    102.362063   \n",
+       "\n",
+       "       Amount_max_fraud  \n",
+       "count      56962.000000  \n",
+       "mean          -0.005195  \n",
+       "std            0.942043  \n",
+       "min           -0.046062  \n",
+       "25%           -0.046062  \n",
+       "50%           -0.046062  \n",
+       "75%           -0.046062  \n",
+       "max           21.709793  "
+      ]
+     },
+     "execution_count": 60,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "X_test.describe()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "그리고 Kearas 모델을 구축합니다."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 61,
+   "metadata": {
+    "_cell_guid": "6c0f7416-412a-4415-94e8-362fa13948ea",
+    "_uuid": "9c648e3b03efea588dea1d20893bf148151447f7"
+   },
+   "outputs": [],
+   "source": [
+    "# fix random seed for reproducibility\n",
+    "np.random.seed(2)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "5 층 네트워크로 마지막은 결과 레이어로 구성하였습니다. 첫 레이어는 Input Dimension을 X_train 칼럼 수인 31로 하였습니다. 마지막 레이어는 softmax 활성화를 사용하였고 나머지는 relu로 활성화하였습니다. 모든 레이어는 0.9인 dropout을 적용합니다. (각 층의 90%의 정보만 사용한다는 뜻입니다.)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 66,
+   "metadata": {
+    "_cell_guid": "bbf4cf25-3052-4b8f-9ac7-700865de4905",
+    "_uuid": "cf449dc135ffd673be7ef9b800df733c6c48831d"
+   },
+   "outputs": [],
+   "source": [
+    "model = Sequential()\n",
+    "model.add(Dense(64, input_dim=31, activation='relu'))\n",
+    "model.add(Dropout(0.9))\n",
+    "model.add(Dense(64, activation='relu'))\n",
+    "model.add(Dropout(0.9))\n",
+    "model.add(Dense(64, activation='relu'))\n",
+    "model.add(Dropout(0.9))\n",
+    "model.add(Dense(64, activation='relu'))\n",
+    "model.add(Dropout(0.9))\n",
+    "model.add(Dense(2, activation='softmax'))  # With 2 outputs"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "loss는 binary cross-entropy로, optimizer는 adam으로 모델을 학습할 수 있도록 설정(compile)합니다. 평가기준(metrics)은 정확도(accuracy)로 받습니다."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 67,
+   "metadata": {
+    "_cell_guid": "2dec00ca-e47c-4142-89bc-e7a445314ebc",
+    "_uuid": "086c950d76895f638b313701200a144fdf3851b6"
    },
    "outputs": [],
+   "source": [
+    "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "컴파일된 모델을 Training 데이터로 학습(Fit)합니다. 일단 batch_size 2048로 두고 10 epoch(세대)만 학습합니다."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 68,
+   "metadata": {
+    "_cell_guid": "bbbd50fb-41e8-4f71-b497-f8e19038dad0",
+    "_uuid": "239658b1175b942a7aa179f8ae8b71dccbcb8f6d"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Epoch 1/10\n",
+      "227845/227845 [==============================] - 5s 20us/step - loss: 0.9254 - acc: 0.8761\n",
+      "Epoch 2/10\n",
+      "227845/227845 [==============================] - 3s 12us/step - loss: 0.1404 - acc: 0.9815\n",
+      "Epoch 3/10\n",
+      "227845/227845 [==============================] - 3s 13us/step - loss: 0.0815 - acc: 0.9913\n",
+      "Epoch 4/10\n",
+      "227845/227845 [==============================] - 3s 14us/step - loss: 0.0603 - acc: 0.9949\n",
+      "Epoch 5/10\n",
+      "227845/227845 [==============================] - 3s 13us/step - loss: 0.0497 - acc: 0.9961: 1s\n",
+      "Epoch 6/10\n",
+      "227845/227845 [==============================] - 3s 13us/step - loss: 0.0447 - acc: 0.9970\n",
+      "Epoch 7/10\n",
+      "227845/227845 [==============================] - 3s 13us/step - loss: 0.0401 - acc: 0.9972\n",
+      "Epoch 8/10\n",
+      "227845/227845 [==============================] - 3s 13us/step - loss: 0.0373 - acc: 0.9977\n",
+      "Epoch 9/10\n",
+      "227845/227845 [==============================] - 3s 13us/step - loss: 0.0356 - acc: 0.9978\n",
+      "Epoch 10/10\n",
+      "227845/227845 [==============================] - 3s 13us/step - loss: 0.0345 - acc: 0.9979\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "<keras.callbacks.History at 0x2ea76744898>"
+      ]
+     },
+     "execution_count": 68,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "epoch = 10\n",
+    "batch_size = 2048\n",
+    "model.fit(X_train, Y_train, epochs=epoch, batch_size=batch_size)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 69,
+   "metadata": {
+    "_cell_guid": "e45c27c0-1e29-47cf-befe-bba155dd21cf",
+    "_uuid": "a6790157aaf0370faa31223ebd4c1faeeaf84f9a"
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "56962/56962 [==============================] - 4s 67us/step\n",
+      "Test score: 0.027575257398611885\n",
+      "Test accuracy: 0.9982795547909132\n"
+     ]
+    }
+   ],
    "source": [
     "score, acc = model.evaluate(X_test, Y_test)\n",
     "print('Test score:', score)\n",
     "print('Test accuracy:', acc)"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "**99.82%의 정확도를 Test에서 얻을 수 있습니다!** 이걸로 충분할까요? 10 epoch 이상으로 적용하면 아마 더 도움이 될 것입니다."
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {
@@ -3189,8 +4430,6 @@
     "_uuid": "343bbc820dd1043196daef8b36ec7b8da445eaec"
    },
    "source": [
-    "** We get 99.82% Accuracy!** Is it good enough for the problem? May be not. More than 10 epochs will still help.\n",
-    "\n",
     "### Training and testing accuracy and loss vs epoch:\n",
     "\n",
     "Let us plot train and test accuracy and loss vs. epoch collcting the history by running the model again:"