The Study Of Quantization Selection Strategy Based On Wavelet Packet Transform

Stock market is a affected by factors such as politics,economy,investor psychology of complex nonlinear systems,its share price movements tend to be stable and non-gaussian white noise disturbance in time sequence,so it is difficult to use traditional methods for de-noising and prediction.Wavelet analysis is a kind of high performance signal analysis method,is developed according to the requirement of the-domain localization,have adaptive features,has been hailed as a mathematical microscope,it has excellent time-frequency localization analysis denoising ability and function,especially suitable for non-stationary nonlinear signal processing,is widely used in signal processing,image compression and speech analysis,and many other fields,the non-stationary time series with good analytical capability.In this paper,on the basis of the research background and the related literature review,according to the proposed research hypotheses-data gathering process-simulation experiment demonstrated in thinking,adopt the method of computer simulation experiment,with the benchmark Shanghai composite index downs a complete cycle as sample data,at the close of the stock price signal wavelet packet transform to analyze its application in the stock market.First of all,this article choose Symlets has disappeared fourth-order moment in the wavelet family "S4" symmetrical wavelet base on the Shanghai composite index closing nonlinear threshold wavelet packet de-noising,wavelet packet transform is studied the significance and effectiveness of denoising;Secondly,on the basis of eliminate random interference,aiming at the shortcomings of the traditional average strategy to buy or sell signal hysteresis,according to the different signal decomposition level of wavelet coefficients can reflect the basic and secondary trend and has no hysteresis characteristics,different decomposition level based on the extraction of low frequency component,to improve the traditional average strategy analysis,and the strategy and strategy of traditional average respectively through simulation experiment,study the advantage and applicability of wavelet quantitative strategies;Finally,the paper also optimizes and improves the analysis of some shortcomings of the wavelet strategy,and studies the practicability and reliability of the optimization strategy.Research shows that:(1)after wavelet packet transform de-noising of the original signal can significantly reduce the noise,compared with traditional wavelet de-noising and non sampling wavelet de-noising,wavelet packet denoising SNR higher,smaller mean square error,has better denoising effect,which can filter out the market significantly noise and random disturbance.(2)by cancel after the noise signal of low frequency component of the building of the simulation experiment,the result proves that under the same level of risk,wavelet quantitative strategies in basic and secondary trend of buy and sell signals at the same time can be shorten the lag of trading signals and has better performance.(3)after the parameter optimization,position control and stop loss setting,the performance of the strategy has been improved significantly.This article research results have certain theoretical significance and practical significance,not only the application of the signal processing method of wavelet packet transform to the financial sector,but also by constructing a set of quantitative trading and back-test framework,has carried on the simulation to the hypotheses,thus promote the results of the study are applied to the practice of investment.