Practice of a Two-Stage Model Using Support Vector Regression and Black-Litterman for ETF Portfolio Selection

Robo-advisor is a hot topic in the field of financial technology (FinTech) in recent years. This study proposes a programmatic ETF portfolio configuration that combines SVR and Black-Litterman's two-stage model. The results of the study showed that the MSE of the first stage of the model was 2.7970 with good performance. According to the prediction result of the first stage, the parameter setting of the Black-Litterman is performed, and then the model is constructed to further adjust the configuration. The final results show that under the same risk value, the SVM+BL two-stage model proposed by this study has a higher return rate than historical return and implied return. Therefore, it can be provided as a reference for the development of an ETF investment strategy.