Research On Financial Data Prediction Model And Trading Strategy Algorithm Based On Fuzzy Theory

With the advent of the information era,to obtain the information we need,people requires more excellent data processing in all aspects,and the contact of the different data characteristic can better help us to analyze and judge,then to take appropriate action.In the financial sector,the circumstances of economic globalization continues to develop,and the financial investment has gradually become an important part of social life,financial data information is playing an increasingly important role.But financial data,including daily stock trading are often affected by many factors,such as human short-term complex psychological factors and the development trend of industry.Accordingly,the daily stock trading data are high dimensional and uncertain,and its characteristic is the inherent nonlinearity,which poses a great challenge to stock trading decisions.Most of the existing methods have focused on a single stock,and the relation of the social network affecting stock trading is rarely considered.(1)In this paper,a fuzzy based stock profit & volatility investment strategy(abbreviated as FSPVIS)is proposed and an investment strategy based on BP neural network is further constructed as the NN-FSPVIS.In the first strategy,the FSPVIS constructs a fuzzy inference mechanism with fuzzy volatility and interest rate for stock trade decision in order to introduce human-operating trading rules and produce the uncertainty of the data.Furthermore,the NNFSPVIS is realized to utilize BP neural network in order to learn model parameters on the basis of the FSPVIS,in which a stock industry relationship mathematical model is built by considering the influence of the stock market related stocks.To the best of our knowledge,this is the first work that combines volatility,interest rate,the related industry stock,buy & sell shares to calculate the stock trading decision value in daily stock trading strategy system.Experimental results on the data of Hongkong stock exchange bank industry stocks verify the effectiveness of the proposed method compared with existing methods.(2)A multi-objective portfolio investment strategy based on fuzzy inference mechanism is proposed.Considering instability of stock data,and using historical returns is not enough to predict future returns,the output variable DA of fuzzy inference mechanism in FSPVIS combines the yield and volatility,and absorbs investment experience of investor.Therefore the paper using output variables DA from fuzzy reasoning mechanism in FSPVIS instead of history yields to predict future earnings,thus to build a new portfolio investment model.Further the fuzzy multi-objective optimization algorithm is used to transform it into a single target linear programming problem,and the effectiveness of the strategy is verified by simulation experiments.