Research on the Price of Stock on AR, MA, ARMA, and ARMA-GARCH Models

This paper aims at issues of Goldman Sachs’ stock forecasting in a short time by using the time series analysis based on four models: AR, MA, ARMA, and ARMA-GARCH models and chooses the optimal model. In this paper, after selecting the sample data and preprocessing data, the regression evaluation index is used to analyze the preliminary models. After that, use the Sequence Stationarity Test and ADF Test to test the series’ stationarity, and analyze the solution of the ARMA model to conclude the formula. The regression evaluation parameters are then compared to the initial models. Later, by selecting from Gaussian distribution, student t distribution, and biased student t distribution, the solution of the AGMA-GARCH model is analyzed. By constructing ARMA and GARCH models, the short-term forecast stock price results are valid and feasible. It concludes that the Arma-GARCH model greatly improves the accuracy of stock forecasting.