Machine Learning for Capital Market Research and Portfolio Optimization

Abstract

Selecting stocks from a large number of active stocks is a critical investment decision. In this study, traditional and machine learning techniques are employed to identify promising stocks. The proposed strategies incorporate historical price paths into momentum techniques and remove stocks with extreme returns. It enhances the fundamental investment decision of stock selection to construct optimized portfolios. These methodologies outperform the standard momentum technique, reduces transaction costs and hedges investors during financial crises.