Support Vector Regression Algorithm under in the CAPM Framework

Abstract

Invertors use the Capital Asset Pricing Model framework (CAPM) to quantify the relationship between return rate and the risk premium; however, CAPM requires many assumptions which are difficult to adapt in the natural business environment. Hence, this paper used an alternative as the Support Vector Regression algorithm (SVR) under the CAPM to explain the relationship between the return rate and the risk premium of the portfolios. This study applied the experimental method based on the data collected in the HOSE market from January 2012 to October 2020. The stock is classified into 10 portfolios, sorted by increasing beta coefficient, and rearranged every six months. The research result proved that the SVR has a better interpretation than the CAPM based on the Wilcoxon test.