Co-movement of precious metals and forecasting using scale by scale wavelet transform

Abstract

In this paper, a new approach is proposed to improve forecasting performances. We analyze the co-movement of precious metals (daily data of gold, silver and platinum starting from July, 2011) using multiple wavelet coherence and determine the movement dependencies on frequency–time space. The data is split into frequencies using scale by scale continuous wavelet transform. All three time series retaining the same frequency scale are (i) selected, (ii) inversed and (ii) forecasted using multivariate model, Vector Auto Regressive Moving Average (VARMA). We conclude that the efficiency of VARMA forecasting is substantially increased because of same frequency highly correlated time series obtained by using scale by scale wavelet transform. Moreover, the direction of price shift (increasing/decreasing trend) is prospected to an adequately distinguishable degree.