Evaluating the Performance of Investment Portfolios Using Deep Learning Algorithms: A Case study of the Cryptocurrency Market (2020-2022)

Abstract

   The research aims to evaluate the performance of the behavior of investment portfolio returns and to help make optimal decisions in conditions of uncertainty as well as high-risk investments. The study was applied on cryptocurrency market whose are characterized by high-frequency trading.

Moreover, we have used Markowitz model based on the concept of return-risk (Mean-Variance) and deep learning based on the work of artificial neural networks algorithms(ANN) and the long short term memory network(LSTM). A random portfolio of 10 assets consisting of cryptocurrencies was selected based on the database of the website: https://finance.yahoo.com/crypto/ during the period 2020-2022 with Python programming.

Finally, we try to evaluate the performance of the models used in accurately predicting the optimal relative weights of the investment portfolio. Which proved the relative effectiveness of the deep learning models by estimating the values of the mean squared error (MSE) at the level of 0.1602% for the deep learning models, compared to a value of 1.3038% for the outputs of the return-risk model for Markowitz, it was based on training 80% and testing 20% of the Dataset.

 Finally, the second hypothesis was accepted, which stipulates the effectiveness of deep learning algorithms by taking advantage of the diversification method in building optimal portfolios with an estimated return of 99%, a risk of 65%, and a value of the Sharpe index estimated at 1.99%. An estimated return of 16.49% and a risk of 12.23%, with no diversification of investment on all portfolio assets, and a Sharpe index value of 0.33%.