Study On Fluctuation Of High-frequency Stock-Price Time Series Based On Multifractal Spectrum

The price behavior is the foundation of Finance. Numerous scholars focus on the study of movement law on financial asset's price, especially on stock-price. And the key point of the study is whether the fluctuation characteristics of stock-price can be grasped and even be forecasted or not. And study on the movement law of high-frequency stock-price time series with multifractal theory is similar to observe the same object with different magnifiers. That's to say, the fluctuation of stock market can be known more truly and more comprehensively through more careful decomposition. And this has the important theoretical and practical significances without doubt.Involving plentiful literatures, collection of high-frequency data and computer programs, positive analyses on high-frequency stock-price time series of Chinese mainland are carried out in this thesis. And the results show that both SSECI and individual stock-price have the obvious multifractal structures and characteristics. And from there different aspects, this thesis forecasts the fluctuation of stock-price with multifractal spectra and its parameters.This thesis has the following five parts.The first part is an introduction, which elaborate the academic and practical significances of applying the multifractal theory to Finance, domestic and overseas relevant literatures, shortcomings on applications, the intents, the contents as well as the innovations of this thesis.The second part is elaborating the fractal and multifractal theory applied in Finance, as well as the multifractal algorithm adopted by this thesis.The third part is forecasting the fluctuation of high-frequency stock-price time series with the shapes and parameters of multifractal spectra. The theoretical anomalous characteristics of multifractal spectra on high-frequency stock-price time series are firstly deduced during prices fluctuate sharply. Then taking four stocks'high-frequency trading data chosen at random as examples, their multifractal spectra's shapes and key parameters are consistent with the above theoretical anomalies. Thence, the start and end of abnormal fluctuation on financial asset price can be predicted by using this method.The third part is forecasting the short-term fluctuation of high-frequency stock-price time series with the parameters of multifractal spectra. Taking Mingsheng...