This study aims to make investment decisions in stock markets using forecasting-Markowitz based decision-making approaches.
The authors’ approach offers the use of time series prediction methods including autoregressive, autoregressive moving average and artificial neural network, rather than calculating the expected rate of return based on distribution.
The results show that using time series prediction methods has a significant effect on improving investment decisions and the performance of the investments.
In this study, in contrast to previous studies, the alteration in the Markowitz model started with the investment expected rate of return. For this purpose, instead of considering the distribution of returns and determining the expected returns, time series prediction methods were used to calculate the future return of each asset. Then, the results of different time series methods replaced the expected returns in the Markowitz model. Finally, the overall performance of the method, as well as the performance of each of the prediction methods used, was examined in relation to nine stock market indices.
