Skip to content

Latest commit

 

History

History
31 lines (21 loc) · 3.17 KB

README.md

File metadata and controls

31 lines (21 loc) · 3.17 KB
Raw

Twitter Bot - Market Trend Prediction using Sentiment Analysis

Based on published reserarch paper

Twitter bot MarketPredGuy

Sentiment Analysis

Monitors a set of tickers: AMD, GOOG, FB, MMM, CAT, AMZN, INTC, PM, MS, JPM, MU, TSLA

Collects sentiment from:

  • Twitter
  • Seeking alpha articles and comments (experimental and in progress)

Machnine learning Technical analysis of stock using 15 technical indicators and sentiment signals.

Based on machine learning SVM (with the RBF kernel) classifier baseline.

In order to construct a decent baseline model, we made use of ten common technical indicators which led to a total of fifteen features as follows.

  • Moving Averages (MA). A moving average is frequently defined as a support or resistance level. Many basic trading strategies are centred around breaking support and resistance levels. In a rising market, a 50-day, 100-day or 200-day moving average may act as a support level, and in a falling market as resistance. We calculated 50-day, 100-day and 200-day moving averages and included each of them as a feature.
  • Williams %R. This indicator was proposed by Larry Williams to detect when a stock was overbought or oversold. It tells us how the current price compares to the highest price over the past period (10 days).
  • Momentum (MOM). This indicator measures howthe price changed over the last N trading days. We used two momentum-based features, one-day momentum and five-day momentum.
  • Relative Strength Index (RSI). This is yet another indicator to find overbought and oversold stocks. It compares the magnitude of gains and losses over a specified period. We used the most common 14 days period.
  • Moving Average Convergence Divergence (MACD). This is one of the most effective momentum indicators, which shows the relationship between two moving averages. It generates three features: MACD, signal, and histogram values.
  • Bollinger Bands is one of the most widely used technical indicators. It was developed and introduced in the 1980s by the famous technical trader John Bollinger. It represents two standard deviations away from a simple moving average, and can thus help price pattern recognition.
  • Commodity Channel Index (CCI) is another a momentum indicator introduced by Donald Lambert in 1980. This indicator can help to identify a new trend or warn of extreme conditions by detecting overbought and oversold stocks. Its normal movement is in the range from -100 to +100, so going beyond this range is considered a BUY/SELL signal.
  • Average Directional Index (ADX) is a non-directional indicator which quantifies the price trend strength using values from 0 to 100. It is useful for identifying strong price trends.
  • Triple Exponential Moving Average (TEMA) was developed by Patrick Mulloy and first published in 1994. It serves as a trend indicator, and in contrast to moving averages it does not have the associated lag.
  • Average True Range (ATR) is a non-directional volatility indicator developed by Wilder . The stocks and indexes with higher volatility typically have higher ATR. The features were all normalised to zero