Stock Market Index Enhancement via Machine Learning

The paper Stock Market Index Enhancement via Machine Learning outlines a general simulation and machine learning framework to formulate optimal portfolios given a pool of stocks, and several industry and style constraints. This group shares a detailed instructions and python code to construct the data set, which are mainly company features used to calcualte the expected returns, as a main input to mean-variance portfolio optimization framework. Moreover, the authors of this paper also share the portfolio construction code in the file "EMEMAR_Portfolio_Choice_Module_Liangliang_Zhang.py" with detailed remarks. The word document is a description on how to use the code to generate data and the main methodology.