Copper Metal Price Using Chaotic Time Series Forecating

Abstract

The present research focuses on the ability to generate a predictive model of the future value of the price of copper using time series of the international price of copper from the last 30 years and the study of chaotic systems. The interest of this study lies in the economic, political and social impacts on a large number of countries with economies based on non-renewable mineral resources. Our approach uses the theoretical foundation of nonlinear dynamic systems for the characterization and analysis of the time series with chaotic component, by using nonlinear tools and methods. Recurrence visual analysis, the Fourier spectrum, the autocorrelation function, mutual information and the Lyapunov exponents are the methods used in the analysis phase of the series, and then a predictive model was generated using nonlinear system techniques. The results identify two cycles of copper prices and price forecasts in the short term. These cycles are connected by a transient and provide further evidence of the effect of investment cycles and infrastructure industries to the value of commodities.