Stock Market Investment Using Machine Learning

Genetic Algorithm-Support Vector Regression (GA-SVR) and Random Forest Regression (RFR) were constructed to forecast stock returns in this research. 15 financial indicators were selected through fuzzy clustering from 42 financial indicators, then combined with 8 technical indicators as input space, the 10-day stocks return was used as labels. The results show that GA-SVR and RFR can make compelling forecasting and pass the robustness test. GA-SVR and RFR exhibit different processing preferences for features with different importance. Furthermore, by testing stock markets in China, Hong Kong (China) and the United States, the model shows different effectiveness.