Effectiveness of OLS and SVR in Return Prediction: Fama-French Three-factor Model and CAPM Framework

Abstract

This study aims to apply the Support Vector Regression (SVR) algorithm under the Fama-French 3-factor model (FF3) theoretical framework to predict the return of a diversified portfolio. Data was collected on the Ho Chi Minh City Stock Exchange (HOSE) from 2010 to 2022 by month, with 145 observations. Firstly, the study compared the two theoretical models, the Capital Asset Pricing Model (CAPM) and FF3, and investigated that FF3 explains volatility in portfolio returns better than CAPM. Furthermore, to forecast the return rate, the study applied Ordinary Least Square (OLS) for CAPM and FF3 and SVR methods for CAPM and FF3. This research used the roll window method to train-test the model and Root Mean Square Error (RMSE) to evaluate the model’s accuracy. An F test was used to compare the performance of these four models. As a result, the FF3 model uses the SVR algorithm more efficiently than the OLS and CAPM.