The application of fractal wavelet to stock investment

Time-varying Hurst index is introduced to characterize the stock stochastic evolution process. The inherent scaling property of wavelet is well suited to the analysis of locally self-similar process. The estimation formula of Hurst index are proposed based on Daubechies wavelet and the consistence of estimation value with true is also proven. Week-index analysis of Shanghai stock market is taken as real example. The result reveals that three types of stock investment strategies are adopted at various stages.