التنبؤ بالسلاسل الزمنية باستخدام النموذج الهجين ARIMA-SVR بالاعتماد على التحويل الموجي المنفصل

Abstract EN

This paper using three models for predicting time series, auto regressive integrated moving average based on discrete wavelet transform DWT-ARIMA model and the support vector regression based on discrete wavelet transform DWT-SVR and hybrid model by combining the ARIMA model with the SVR model based on discrete wavelet transform DWT-ARIMA-SVR, these models are used to predict monthly global gold prices, based on a time series of monthly global gold prices, the time series data was used during the period from Jan-1991-Dec-2021, to compare hybrid model of DWT-ARIMA-SVR with individual models DWT-ARIMA and DWT-SVR, these models are compared using the MSE, MAE and MAPE,T prediction accuracy measures to find the most appropriate model for predicting future values.

this paper reveals the superiority of DWT-ARIMA-SVR on other individual models by having the lowest values of the prediction measures.