Skip to content

Commit

Permalink
update figure
Browse files Browse the repository at this point in the history
  • Loading branch information
zhimin-z committed Sep 25, 2024
1 parent 15527bf commit 98a901d
Show file tree
Hide file tree
Showing 5 changed files with 110 additions and 51 deletions.
2 changes: 1 addition & 1 deletion docs/catalog.md
Original file line number Diff line number Diff line change
Expand Up @@ -46,4 +46,4 @@

[詹好](https://github.com/zhanhao93)负责了项目的初期规划与统筹,并参与了第一版的编辑和审核;[赵志民](https://github.com/zhimin-z)主导了项目二期的更新与维护,并负责全书最终编辑和校验;[李一飞](https://github.com/leafy-lee)参与了第1-5章内容的编辑;[王茂霖](https://github.com/mlw67)参与了第2-6章内容的编辑。

另外,特别鸣谢了[谢文睿](https://github.com/Sm1les)[杨昱文](https://github.com/youngfish42),他们共同提供了本书的在线阅读功能;[张雨](https://github.com/Drizzle-Zhang)对第2章的早期内容进行了修订各成员的协作确保了本书高质量的编写和顺利完成。
另外,特别鸣谢了[谢文睿](https://github.com/Sm1les)[杨昱文](https://github.com/youngfish42),他们共同提供了本书的在线阅读功能;[张雨](https://github.com/Drizzle-Zhang)对第2章的早期内容进行了修订各成员的协作确保了本书高质量的编写和顺利完成。
2 changes: 1 addition & 1 deletion docs/chapter6.md
Original file line number Diff line number Diff line change
Expand Up @@ -69,7 +69,7 @@ $$
**122页**介绍了一种将样本空间划分成多个互不相容区域的方法,然后对各区域内的正例和反例分别计数,并以多数类别作为区域中样本的标记。这种方法本质上不同于参数方法,它并不是在参数空间中进行搜索构建划分超平面,而是在泛函空间上直接进行搜索。一个典型的例子是我们熟悉的决策树模型:

<div style="text-align: center;">
<img src="images/decision_tree.png" alt="decision_tree" width="300" height="320"/>
<img src="images/decision_tree.png" alt="decision_tree" width="300" height="480"/>
</div>

每当构造一个决策树的节点时,相当于在样本空间上进行了一次划分(即划分机制)。这种洞察方式同样适用于解释剪枝操作,即通过减少不必要的节点来简化树结构,同时保持或提高模型的性能。
Expand Down
6 changes: 3 additions & 3 deletions docs/chapter8.md
Original file line number Diff line number Diff line change
Expand Up @@ -134,11 +134,11 @@ $$
- 更新:
$$
\begin{equation}
\begin{align*}
\begin{align*}
\Sigma_{t+1} &= \Sigma_t + \nabla l^{\eta}_t(x^{\eta,l}_t) \nabla l^{\eta}_t(x^{\eta,l}_t)^\top \\
x^{\eta,l}_{t+1} &= \Pi^{\Sigma_{t+1}}_D (x^{\eta,l}_t - \frac{1}{\beta} \Sigma_{t+1}^{-1} \nabla l^{\eta}_t(x^{\eta,l}_t))
x^{\eta,l}_{t+1} &= \Pi^{\Sigma_{t+1}}_D (x^{\eta,l}_t - \frac{1}{\beta} \Sigma_{t+1}^{-1} \nabla l^{\eta}_t(x^{\eta,l}_t)) \\
\end{align*}
\end{equation}
\end{equation}
$$
其中 $\nabla l^{\eta}_t(x^{\eta,l}_t) = \eta g_t + 2 \eta^2 g_t g_t^\top (x^{\eta,l}_t - x_t)$

Expand Down
Binary file modified docs/images/decision_tree.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Loading

0 comments on commit 98a901d

Please sign in to comment.