Nonparametric Indentification,Path Analysis And Frequency Mixing Prediction Of Financial Index

With the continuous deepening of reform and opening up,the pattern of financial market is changing,the tendency of mutual promotion and mutual development is indicated between finance and economic development.Financial institutions have worked out different financial indexes to reflect the overall price of the securities market or the financial index of a certain stock with its price fluctuation or trend by implementing different descriptive statistics methods.Financial index is not only the comprehensive indicator to measure the trading fluctuation range and business condition for the whole market,but also the important basis for investors to make decisions.After doing an in-depth and systematic research on financial index,which help us better realize the fluctuation rules of economic phenomenon and provide some reliable basis to establish the macroscopic policies.CSI Well-Off was created by China Securities Index Co.Ltd in 2005.It aims at reflecting the overall situation of the listed companies which comprehensively represent the orientation of the industrial structure optimization of national economy.In particular,this index is not the arithmetic product of float and its price but weighting the comprehensive financial indexes,which could really reflects the national economic development.If we treat the stock market as the barometer of Chinese economy,then CSI Well-Off would indicate the development of our economy has got into the New Normal.CSI Well-Off behind massive trading data is of discernible,regular information,which has higher research value and market value,but it is difficult to directly support the policy decision,that is why CSI Well-Off is just a statistical description,which cannot make the inferential statistical analysis.Even analyzing with technical indicators,we only can do some scenario analysis and subjective anticipation of the potential.The study reconstructed CSI Well-Off by using the pure technical indicators based on the concepts of big data analysis and transferred the descriptive statistics to the inferential statistics and then imported the classic scenario analysis into the verifiable simulation prediction within big data.The main innovation and contributions of this paper are as followed:Firstly,refactoring CSI Well-Off.Based on "what is",the analytical principle,which the big data focused on,introducing the descriptive statistics index of well-off into the inferential statistical index of well-off with pure technical indicators,we have reconstructed CSI Well-Off and made it possible of decision support.The index of well-off reconstructed by technical indicators,is just a virtual index value of "what is" for real results,but cannot describe the reason of "why" for the index’s structure.However,the refactoring CSI Well-Off could capture the linear and nonlinear data correlation between the technical indexes and the index value and reveal the tendency of the index change and operation rules,it is of great decision support significance to monitor financial markets,prevent and control financial risks and maintain the exponential stable operated.This study selected CLOSE,EXPMA,ROC,SRMI,BIAS,MTM,SI and WVAD as observation variable based on dynamic factors and established REAL-TIME model.Using mixed frequency of weekly data and monthly data to predict CSI Well-Off between January 2015 and December 2015.Besides,I analyzed the predicted effect of extra-sample and verified the robustness of the model by using the Bootstrap method.Secondly,identifying the important technical indicators.This study finds out Reverse trend technical index after using non-parametric Random forests method.SI,RSI,BIAS,DPO and CCI are the most significant technical indicators to confirm CSI Well-Off,then are CDP,VR,KDJ,MTM and ART.This suggest that the tendency of CSI Well-Off is not obvious and most of it mainly remain consolidation,where existed many reversal opportunities.Thirdly,finding out the variation of the technical indicators decided fluctuations.It is the first time to use time-varying coefficient of path to analyze and find that from July 8th in 2005 to February 27th in 2009,CCI,DMI,POC,RSI and WI determined tendency,white WVAD,TRIX,VOSC,BBI caused fluctuations;From March 6th in 2009 to March 20th in 2015,VOSC and CCI determined tendency,white EXPMA,PRICEOSC,WVAD caused fluctuations;From March 27th in 2015 to December 25th in 2015,BBI,BIAS,DMA,DPO,MIKE,PVT determined tendency,white STD and SAR caused fluctuations.Comparing with the influence of the same variable in different ranges and finding that the influential effect is alike,but the impact degree is of different size.From July 8th in 2005 to February 27th in 2009,VOSC had an impact on CSI Well-Off is much more than that from March 6th in 2009 to March 20th in 2015,which indicates CSI Well-Off is more vulnerable to the impact of fluctuant volume.