Naive Bayes Classification for Multi-Factor Stock Selection

Summary

The article explains naive Bayes classification and applies it to a stock selection strategy. The classifier estimates the probability of each outcome from historical labeled observations, assuming features are conditionally independent and equally important. The predicted class with the largest probability determines the classification. It also describes standardizing each feature by subtracting its mean and dividing by its standard deviation.

For the strategy, nine technical indicators serve as features. A stock is labeled positive when its return over the next 22 trading days exceeds 5%; otherwise it is labeled negative. The model is trained on historical data, then buys when the prediction is positive and no position is held, and sells when negative while holding a position. The article reports a daily backtest period from September 2015 through March 2017 and initial capital of 10 million, but the provided text includes no readable performance results. The independence assumption, feature selection, and limited backtest evidence constrain what can be concluded about predictive value or robustness.

Key ideas

• Naive Bayes assigns an item to the class with the highest estimated conditional probability.
• The method assumes features are conditionally independent and equally important.
• Nine technical indicators are used as features for a stock classification model.
• The positive label is defined by a return threshold over the next 22 trading days.
• The article describes a daily backtest but supplies no readable performance figures.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.