Time Series Analysis for the COMEX Copper Spot Price by Using Support Vector Regression

Abstract

In this research work, support vector regression (SVR), a powerful machine learning tool that can identify nonlinear patterns effectively thanks to the introduction of a kernel function is used. Three different strategies (direct multi-step scheme, recursive multi-step scheme and direct-recursive hybrid scheme) for automatic lag selection in time series analysis are proposed. This article examines the forecasting performance of the three kinds of SVR models using published data of copper spot prices from the New York Commodities Exchange (COMEX). The numerical results obtained have shown a better performance of the direct-recursive hybrid scheme than the recursive multi-step scheme and direct multi-step scheme. The findings of this research work are in line of with some previous studies, which confirmed the superiority of SVR models over other classical techniques in relative research areas.