The Comparison of Prediction and Optimization Algorithms in Portfolio Selection on Tehran Stock Market

In this study, two algorithms, neural networks and fuzzy neural networks, are used to predict the stock prices, and two memetic algorithms, particle swarm and genetic algorithm, together with quadratic method are used to solve unbounded optimization problem by applying daily data of 23 listed companies on Tehran Stock Market during 2012-2015. The findings show that artificial neural networks have better performance than fuzzy neural network in predicting yields. In addition, the comparison of performances of quadratic, genetic and shuffle frog Leaping (SFL) algorithms reveal that SFL algorithm has recorded better performance than genetic and quadratic algorithms. As a result, artificial neural network may be a reliable algorithm for shareholders.