Comparison of Predicted Tehran Stock Exchange Cycles using ANFIS, MLP, RBF and PNN Networks based on PSO Algorithm

One of the financial institutions of the countries in the world is the Stock Exchange, which is used as indicators of economic health. As for a large volume of capital is managed through the stock market and is placed in the production and industry, the recession in this market can have important effects. In this paper, we seek to extract the cycles of prosperity and recession in the Tehran Stock Exchange using the TEPIX and the Pagan and Sossounov method. Then, using the PSO algorithm, the most important predictor variables are determined and in the next step We predict market financial cycles using ANFIS, MLP, RBF and PNN networks. The results show that according to the mean square error, the root mean square error, the accuracy of the model and the Kappa coefficient, the MLP model provides better results in predicting future market conditions.