The Stock Index Forecast Based On The Portfolio Model

As a representative of high-risk and high-yield financial markets,the stock market has attracted more and more investors' attention since its birth.However,due to the volatility of the stock market,it is particularly important to study the stock market and grasp the overall trend of the market.Besides,as a comprehensive indicator,the stock index reflects the overall trend of stocks and it is also an important factor affecting stock prices.Therefore,the research on stock index forecast has certain research value.We combine two major stock analysis methods(technical analysis and fundamental analysis)to analyze stock index.First,regardless of other factors,we only rely on market performance and forecast based on historical trading data;next,we modeling based on economic indicators and social indicators data,and ultimately we weight the two methods to achieve more accurate predictions.This paper firstly uses the time series method and the neural network method to predict the daily closing price of the next month based on the historical trading data from 2010 to 2018,and then choose the method with better prediction effect.Next,we select the representative 26 economic indicators which affect the stock index and use the monthly data to establish prediction model to forecast the average closing price of the next month.Finally,the predicted daily closing price and the average closing price of the next month are weighted by the average absolute percentage error reciprocal method to obtain the final forecasting result.In this paper,the linear model is effectively combined with the nonlinear model to establish a complete stock index forecasting model.In the end,the empirical research is carried out to prove that the stock index forecast based on the combined model is meaningful.