Translated Nigeria Stock Market Prices Using Artificial Neural Network for Effective Prediction

This paper used error back propagation algorithm and regression analysis to analyze and predict untranslated and translated Nigeria Stock Market Price (NSMP). Nigeria stock market prices were collected for the periods of seven hundred and twenty days and grouped into untranslated and translated train, validation and test data. A zero mean unit variance transformation was used to normalize the input variables in order to allow the same range which makes them to differ by order of magnitude. A 5-j-1 network topology was adopted because of five input variables in which variable j was determined by the number of hidden neurons during network selection. The untranslated and translated data served as input into the error back propagation algorithm and regression model which were written in Java Programming Language. The results of both untranslated and translated statements were analyzed and compared. The performance of translated NSMP using regression analysis or error back propagation was more superior than untranslated NSMP. The results also showed that percentage prediction accuracy of error back propagation model on untranslated NSMP ranged for 11.3% while 2.7% was for translated NSMP. The 2.7% percent accuracy as against 11.3% indicates the relative stability of translated NSMP prediction as against untranslated NSMP. The mean relative percentage error was very low in all hidden topologies of error back propagation of translated NSMP than untranslated NSMP. This indicates that translated NSMP prediction approach was superior to untranslated NSMP predicition.