Forecasting the Price of Crude Oil Using Regression Techniques and Time Series Using Sarima

Abstract

In this paper, we proposed a new time series analysis method for the future prediction of crude oil price, named Seasonal Autoregressive Integrated Moving Average(SARIMA) which is extension of ARIMA.This study aims to upgrade the efficiency of forecasting using time-series, which would thus increases the exactness and reduces the RMSE value of the predictions. The RMSE value is compared with the other previous predicted models. The RMSE value of this method is less.The numerical outcomes are compared with the past techniques. The results of the proposed strategy have demonstrated an increase in the exactness of the crude oil price forecasts. The current crude oil price can be predicted by using the regression techniques. In regression techniques we use two 1.Linear Regression 2.Random forest regressions. In this paper, we find the results of both regressions and then compare the results and tell which is best regression technique for current crude oil price prediction based on the RMSE value. We obtained that the RMSE value of Random Forest is better than the Linear Regression and other cited models.