A Factor Pricing Model Based on Machine Learning Algorithm

This paper adopts Wavelet transform and Support Vector Regression (SVR) algorithm to predict all stock returns and form a time-varying Machine Learning (ML) factor added in CH4 of Liu, Stambaugh, et al. (2019) in the Chinese stock market from 2000 to 2020. The result shows that the decile sorted portfolios formed by the predicted return rate can obtain significant excess returns in the Chinese market. Furthermore, the introducing a ML factor can significantly improve pricing power. Through GRS test, we found that the five-factor model with only ML factor added can pass the test at the 5% significance level. In addition, we also found that our factor model is macro-state dependent, and the better the macroeconomic and market performance, the stronger the pricing power of the factor model