Predictive study on time series modeling and comparison with application

Abstract

Efficient time series modeling and forecasting are essential in different practice areas. Consequently, much active research work on this topic has been ongoing for several years. Given the importance of different prediction methods, this research aims to provide a brief description of some common time series prediction models used with their salient features. Therefore, Box-Jenkins and exponential booting models were compared, along with the strengths and weaknesses of the prediction. Our discussion on various time series models is supported by giving the experimental prediction results, which were made to the actual monthly sales of some fuel products for the period 2014-2017. While installing the Data Set template, special care is taken to select the most creative. To evaluate prediction accuracy in addition to comparing it, we used several criteria, mean square error (MSE), mean absolute deviation (MAD), mean absolute percentage error (MAPE), and mean square error (RMSE). To obtain originality and clarity in our discussion on modeling and forecasting a time series, we were able to obtain assistance from various published research work from famous magazines and some standard books and it was concluded that the 3ARMA terms best model among the Box-Jenkins models built based on the dependence of gas oil sales in Iraq, as well as Simple exponential smoothing is the best exponential smoothing model to forecast in the coming years for sales of improved gas and gas oil in Iraq.