Predicting the Traded Value Index for the Iraqi Stock Market Using Artificial Neural Networks

Abstract

Prediction is one of the important topics in statistical sciences to help leaders develop future plans and make appropriate decisions for them. This research includes one of the modern prediction methods represented by artificial neural network (ANN) models, specifically the multi-layer network, as the back propagation (BP) algorithm was adopted and trained several times to get the lowest error value.

The research aims to predict an important indicator of the Iraqi Stock Exchange, which is the traded value index, using one of the scientific methods, which is artificial neural networks, applying the back propagation algorithm and using the (Matlab 2018) program.

The researcher concluded that the artificial neural network model gives accurate results by spreading the data in the regression line, and the value of the coefficient of determination reached (R = 0.99928), and the mean square error reached (0.0000001), which is very small, and this indicates the accuracy of the prediction. Daily data was collected for the traded value index for the period from January 2, 2018, until October 31, 2021.