Prediction of Stock Price using Particle Swarm Optimization Algorithm and Box-Jenkins Time Series

The purpose of this research is predicting the stock prices using the Particle Swarm Optimization Algorithm and Box-Jenkins method. In this way, the information of 165 corporations is collected from 2001 to 2016. Then, this research considers price to earnings per share and earnings per share as main variables. The relevant regression equation was created using two variables of earnings per share and price to earnings per share, and stock prices were predicted through particle swarm optimization algorithm in MATLAB. IBM SPSS was used to predict stock prices with Box-Jenkins time series. The Results indicate that particle swarm optimization algorithm with 4% error and Box-Jenkins time series with 19% error, have the potential to predict stock prices of companies. Moreover, PSO algorithm model predict stock prices more precisely than Box-Jenkins time series. Also by using EViews 7 software, the results of Wilcoxon-Mann Whitney statistics showed that PSO algorithm predicts the stock price more accurately