股票价格的一种线性分形预测方法

A LINEAR FRACTAL PREDICTION METHOD FOR STOCH PRICE

主要研究股票价格的变化规律及对其未来价格的预测方法．通过对股票市场价格进行小波分析，发现股票价格的变化服从自相似性，并且满足某种拟周期性．而自相似是分形的一个重要特征，就可将股票价格的变化作为分形来进行研究．利用这一结论，找出了股票价格变化的线性分形插值函数，并提出一种线性分形预测方法来对股票价格进行预测．在数值试验中，对四川长虹的股票价格进行了分析，利用上述线性分形预测方法进行预测，得到了一些令人满意的结果，说明该方法是可行的．

Studies the transient law of stock market price and the prediction method of the future stock price. Wavelet transform is used to find the evidence of self-similarity of a function. This method can solve the major problems of the previous analysis in this area. It doesn't depend on the assumption of stationarity of the time series, it can detect structures in data that are highly localized in time and therefore non-detectable by Fourier transform, and it doesn't require very large numbers of observations. There is an equivalent proposition to ensure that this method is feasible. This article uses wavelet transform to analyze the stock market price, and gets some evidence of self-similarity. During the process, some limited evidence of quasi-periodicity in the occurrence of large amplitude shocks to the system is found. Self-similarity is a primary feature of fractal, so the stock market price can be illustrated as fractal. With this conclusion, we find out the linear fractal interpolation function of the stock price, and get a linear fractal prediction method to infer the future stock price. By numerical examples we analyze the stock price of Sichuan Changhong, use the linear fractal prediction method to predict future price, and get some interesting results, which demonstrate that this method is feasible.