The Analysis and Forecasting of Time Sequence Based on SARIMA Model

Abstract

This paper focus on the modeling of quantities of commodity barcodes registrated in Guangzhou. With analysis of the sample time sequence, difference method is applied to turn the sample sequence into stationary one. By studying the autocorrelation function(ACF) and partial autocorrelation function(PACF) figures, seasonal autoregressive integrated moving average (SARIMA) models are put forward. After simulation, SARIMA(0,1,1)(0,1,1) 12 model is proved to be of better accuracy, and the result demonstrates the goodness of fit. By utilizing this model, the paper forecast the quantities of commodity barcodes registration in the next six months. The result indicates the quantities of commodity barcodes registration will develop in the future, with year-on-year growth of about 10%. 2015 18-19, December, ISCC