Predicting Date Production in Iraq Using Recurrent Neural Networks RNN

Abstract

Artificial intelligence methods play an important role in predicting future values of time series and thus help in setting economic and social development plans. The study aimed to predict the production of dates in Iraq using recurrent neural networks, based on the production of dates in Iraq for the period from 2002-2021. The appropriate prediction model was chosen based on the MSE, MAPE, and MAE error measures. Recurrent neural networks that used the TRAINBR training function and the Purlin function were adopted to predict the production of dates in Iraq, which gives the lowest error value for the MSE, MAPE, and MAE error measures.