Algorithm of Hybrid Analysis Systems in Predicting Optimal Portfolio

Abstract

The development of analytical techniques in the investment portfolio in accordance with market conditions and economic climate can increase the dynamism of investment development and achieve greater returns against risk control. Forecasting and optimization techniques help financial decision makers to place the best stocks in their portfolio based on market information and achieve greater returns by optimizing it. The purpose of this research is predicting the effectiveness of the difference between Sortino and Markowitz portfolios is based on the hybrid analysis systems algorithm. Accordingly, 102 companies of Tehran Stock Exchange were examined in the 2014-2018 period. In this study, by separating value stocks and growth stocks, random portfolios were selected to test the research hypotheses and for analysis, two algorithms Support Vector Machines (SVM) and an Adaptive Neuro-Fuzzy Inference System (ANFIS) were used to select the most desirable portfolio. The results showed that Sotino (X) portfolio based on meta-heuristic algorithm (support vector machine algorithm) is significantly different from Markowitz (Y) portfolio, so that decision makers in Sortino portfolio seek to optimize their stock portfolio through long-term growth stocks. It was also found that the accuracy of the adaptive neural fuzzy inference analysis (ANFIS) system is higher than the accuracy of the support vector machine analysis (SVM) system to select the most effective portfolio from Sotino and Markowitz portfolio, because it has two learning mechanisms and Neural network optimization and linguistic expression of fuzzy inference help managers to make better estimates of uncertainty and uncertainty.

Keywords

• algorithm of hybrid analysis systems
• portfolio utility optimization
• support vector machines (svm)
• adaptive neuro-fuzzy inference system (anfis)