MONITORING AUTOCORRELATED PROCESSES WITH WAVELETS

This paper develops a wavelet control chart for monitoring autocorrelated processes. The procedure uses the discrete wavelet transform of the original series, and traditional control charts are applied to the stream of wavelet coefficients. Unlike other control charts for monitoring autocorrelated processes found in the literature, the wavelet control chart does not require that a model be specified for the process data. The wavelet-based control chart is simple enough that it can be easily automated. Real and simulated data are used to illustrate the effectiveness of the proposed wavelet control chart.