Study on the WCC Method for Time Series Data Analysis

In this paper, based on reviews of the wavelet cross-correlation methods (WCC) used for time series analysis briefly, the wavelet cross-correlation degree (WCCD), which can be used to describe the cross-correlations of time series data in the whole time domain, has been defined, and meanwhile the method of drawing wavelet cross-correlation coefficients contour map has also been proposed, by which the integrated time-frequency analyses of cross-correlations of time series can be realized. By applied to analyze two observed time series, the superiority and effectiveness of WCC methods have been verified. Analyses results show that both global and local cross-correlations of time series can be analyzed accurately by WCC methods. Compared with traditional cross-correlation analysis methods, the WCC methods are more applicable and flexible since they can describe the cross-correlations of non-stationary time series under any time scales and any time delays effectively.