Prediction And Spectrum Analysis Of NAV To The Open-end Funds Based On Wavelet Theory

This thesis discusses the forecast for the net assets value of funds and the fluctuation characteristics in time-frequency domain.Employing wavelet analysis,we present empirical studies for the open-end funds in China and show the results of the different open-end fund investment types in-depth analysis.First,the theory of wavelet analysis is introduced and then combining with the ARMA-GARCH model,we establish a new prediction model for the net value of fund.Empirical results show that the wavelet multi-resolution analysis theory improve the utilization rate of the time series by decomposition,so that the wavelet decomposition theory and classical ARMA-GARC H model can be combined to promote the prediction accuracy of the model.Secondly,we introduce the wavelet power spectrum analysis theory.By using the wavelet power,wavelet cross power and wavelet coherency to present the fluctuation energy distribution of the time series in the time-frequency domain,we show a new mining method of analysis for the distribution characteristics of the net value of the funds and the correlation of the funds and the stock market.The empirical results of the 12 open-end funds and the Shanghai Composite Index present that the wavelet power spectrum analysis method is suitable to the analysis of the distribution characteristics of the funds and the correlation of the funds and the stock market,and provide the scientific reference in the practical application for the investors' investment decision,the fund performance evaluation,and the supervision and administration of fund.To sum up,there are three innovations of this thesis as follows:(1)We establish the prediction model of the NAV combining the wavelet decomposition method which could increase the utilization rate of data and the ARMA-GARCH model.Hence,this model improves the prediction accuracy of the model.(2)Using the theory of wavelet spectral analysis,we analyze the time-frequency domain perspective of fund volatility and the correlation between fund and market.(3)Based on the complex wavelet spectral analysis,this thesis gives the concept of wavelet spectrum which provides a new,more complete,with time frequency resolution index to measure the system risk of fund.