Portfolio optimization based on return prediction using multiple parallelinput CNN-LSTM

The success of any investment portfolio always depends on the future behavior and price events ofassets. Therefore, the better one can predict the future of an asset, the more profitable decisions can bemade. Today, with the expansion of machine learning models and their advanced sub-branch i.e. deeplearning, it is possible to better predict the future of assets and make decisions based on thosepredictions. In this article, a deep learning method called CNN-LSTM with multiple parallel inputs isintroduced and is shown that it is able to provide a more accurate prediction of asset returns for the nextperiod than other machine learning and deep learning models. Then, these forecasts will be used in twostages to build the portfolio. First, the assets that have the highest predicted return are selected, and thenin the second step, Markowitz's mean-variance model will be used to obtain the optimal ratio of theselected assets for trading in the next period. The model test is performed on the assets randomlyselected from different New York Stock Exchange industries based on the ۱۱ Global IndustryClassification Standard (GICS) Stock Market Sectors.

Keywords:

portfolio optimization , return prediction , multi-parallel input , mean-variance model