Open AccessForecasting the Accounting Profits of the Banks Listed in Iraq Stock Exchange Using Artificial Neural Networks
Author(s) -
Zahra Hasan Oleiwi Alaameri,
Mustafa Abdulsahib Faihan
Publication year - 2022
Publication title -
webology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.259
H-Index - 18
ISSN - 1735-188X
DOI - 10.14704/web/v19i1/web19177
Subject(s) - stock exchange , artificial neural network , autoregressive model , econometrics , stock (firearms) , computer science , artificial intelligence , economics , finance , engineering , mechanical engineering
This paper demonstrates the feasibility of using deep learning approaches in time series forecasting of bank profits. Two types of neural networks were used, LSTM (Long-Short Term Memory) and NAR (Nonlinear Autoregressive) networks, for comparison. The data from 12 Iraqi banks, which are registered in the Iraq stock exchange, were involved in this study for sixteen years (2004-2019). RMSE and MAPE were used for comparing the performance of the two models (LSTM and NAR). Our results showed that the NAR is more accurate than LSTM for the prediction of profits. And that the use of the NAR network by the Iraqi banks will help them predict future accounting profits.