U.S. unemployment rate prediction using time series model

Abstract

Although previous studies have given a better prediction model for the Americans unemployment rate, due to the short time and different time nodes, the parameters of the model and the seasonality and the stability of the time series are also different. In this study, the ARIMA model, which is the most widely used in the time series, is adopted and the seasonal influence is added to the model according to the selected time period. At the same time, two models are used to predict the unemployment rate in the United States from January 2017 to January 2019. The stability of the model was determined by Dickey-Fuller test, and the fitting and prediction effects of the two models were compared by comparing the values of AIC and MSE. With the fitting prediction method of the unemployment rate in the United States, this paper can analyze and predict the unemployment rate in other Western countries, and can further compare and analyze the reasons with China s unemployment rate, which is convenient for us to better regulate macroeconomic policies.