WAVELET TRANSFORMS FOR THE STATISTICAL ANALYSIS OF RETURNS GENERATING STOCHASTIC PROCESSES

Abstract

We study high frequency Nikkei stock index series and investigate what certain wavelet transforms suggest in terms of volatility features underlying the observed returns process. Several wavelet transforms are applied for exploratory data analysis. One of the scopes is to use wavelets as a pre-processing smoothing tool so to de-noise the data; we believe that this procedure may help in identifying, estimating and predicting the latent volatility. Evidence is shown on how a non-parametric statistical procedure such as wavelets may be useful for improving the generalization power of GARCH models when applied to de-noised returns.