Comparing Black-Litterman Model with Markowitz Model and generating investors' views with ARIMA Model and SVR
This research project is to compare Markowitz Mean Variance Optimizer with Black-Litterman Optimizer. The input of Black-Litterman Model (BL Model as abbreviation) are market equilibrium returns and invesotrs' views. Market capitalisation for ETF and open interest for futures were used as the weight of market returns. The trading signals were generated by ARIMA model and then, used them as featurs of SVR model to generate returns. The results are the ML-generated returns are 5% more return for each component in the portfolio than using futures price to calculate forecasted returns.