A Multi-Factor Rotational Stock Selection Strategy using Machine Learning Models: A Study based on the U.S. Stock Market

This study delves into the advantages and disadvantages of multifactor stock picking strategies formed under different models using a dataset containing the closing prices of ten U.S. stocks and technical indicators such as Simple Moving Averages (SMAs), Exponential Moving Averages (EMAs), Relative Strength Indices (RSIs), Bollinger Bands (BBs), Exponentially Smoothed Moving Averages with Differences and Divergences (MACDs), and returns. The study aims to train and predict the data with different models to derive the economic indicators of different stock picking strategies and make a preliminary judgment of their strengths and weaknesses, and then compare the underlying with the maximum predicted return selected by the machine learning model with the underlying with the true maximum return in order to back test the strategies. The best predictive model is selected based on the scoring results. The final result of this study provides constructive suggestions for investors for future stock selection and investment in the stock market.