Research On Financial Time Series Forecasting Based On Fuzzy Time Series And Grey Theory

Financial time series is an extremely important data type in the economic field.In order to be able to better and faster adapt to the new economic environment brought about by financial globalization,it is therefore necessary to establish a suitable model for predicting and analyzing financial time series.The economic development trend,on the other hand,can reduce investment risks.This paper first introduces the development process of stock prediction theory and forecasting model.The early forecasting models mainly include ARMA,ARIMA,ARCH,etc.With the development of computers,artificial intelligence prediction methods have also been added,mainly including neural networks,genetic algorithms,and support vector machines.These methods have been applied to the prediction of the stock market.This article focuses on the basic theories,basic methods and prediction steps of the grey prediction theory and fuzzy time series.At present,the study of fuzzy time ordering to improve prediction accuracy mainly involves the division of domain sub-intervals and the establishment of a reasonable logical relationship matrix.This paper focuses on the problem of artificially dividing intervals,and first classifies them by the mean of data fluctuations,and then classifies them.The domain partition is divided into two to fit the distribution of data;the study of grey prediction focuses on the ashing of original data and the transformation of background values.This paper introduces two new ashing methods by analyzing the data.To improve the prediction accuracy.The paper introduces the implementation process of two improved models,and then gives the corresponding experimental results and makes an analysis.Finally,the two improved models are combined to propose a grey-fuzzy time series forecasting model.Combined with the impact of public information on stock prices,the daily closing price of the stock and the five-minute time-sharing price are used as experimental data to analyze the forecast results.This confirms that the model performs well in both price forecasting and trend analysis.