The Study Of Stock Market Trends Prediction And Relative Efficiency Measurement Base On Bayesian Method

Effective stock market is very important to the development of modern economy.It provides a convenient place and way for the effective allocation of resources and the fair distribution of social wealth.The securities market efficiency issue has been a research hotpot of debate,the classical finance efficient market hypothesis believe that the market is effective,and the behavioral finance theory think that the market is an ideal state,and the stock market is inefficient or even invalid state,self-similarity and long memory is the futures of the stock market.In addition,the empirical debate is to test whether the market is effective.This paper argues that market efficiency is relative,and market efficiency is a problem of degree rather than a simple,effective and ineffective one.In order to predict the trend of securities market and the market efficiency more accurately,the study based on the single model to predict the trends and then adopted the multi model Bayesian Comprehensive Decision Method.The comprehensive decision process is realized by the judgment of historical trends,the prediction result of models and the real fluctuation of the market.On this basis,the relative efficiency of the market is quantified based on Bayesian statistical ideology.In this paper,we selected the Shanghai Stock Exchange's closing price of Shanghai Composite Index as the research samples,and three mainly research issues:Firstly,the ARMA model and neural network model are established to forecast the Shanghai stock index return series,and the prediction ability under different situations is analyzed statistically according to the prediction results.The results show that the effect of single model prediction is not significant,because the market is relatively effective,the market itself is difficult to predict;in addition,there are many factors affecting the securities market,it is difficult to use a single model to predict it accurately;Secondly,based on a single model the combination model to predict market trends.The two models are integrated by Bayesian method,and the prediction results are fused and analyzed.It is found that the integration method improves the prediction ability,especially when the two models diverge on the market trendprediction,and gives the scientific probability calculation.Thirdly,study the relative efficiency of the market.Based on Campbell's relative market efficiency evaluation idea and model prediction results,combined with model prediction data,the paper calculates the weak relative efficiency of Chinese Stock Market from the perspective of market predictability by Bayesian method.The study shows: first of all,the efficiency of China's securities market is not completely effective or absolute invalidity,but there is a relatively stable relative efficiency;Secondly,because the stock market is changing the complex nonlinear system,single model is difficult to fully reflect market fluctuations in the market prices and market forecast;two model of actual volatility trend of integration base on Bayesian method continuously adjusted according to market model,the prediction effect of combination model has more advantages.Therefore it has important theoretical value and practical significance to study the model prediction method and relative efficiency measurement method of stock trend prediction and market efficiency measure,provide methods and empirical basis for the rules of securities market,formulate regulatory policies and market investment strategy etc.