Applying Machine Learning for Portfolio Switching Decisions

Abstract

Portfolio switching is an investment strategy that responds to Market momentum by adjusting a portfolio to produce more value. This study will use machine learning algorithms to predict the prudent state of the market using sample mutual fund returns to make portfolio-switching decisions. The study trained and tested the performance of sample funds using monthly returns of mutual funds, market proxy, and risk-free assets over seven years. The machine learning algorithm, specifically Support Vector Machine (SVM) and Logistic Regression (LR), was used to select a portfolio that could adapt to a changing market. SVM outperformed LR in terms of performance and evaluating the algorithm’s efficacy using sample mutual funds helps the investors to choose the suitable algorithm for investment decisions.