Fitting and Prediction of Total Retail Sales of Consumer Goods Based on Consumption Indicators

This paper uses eight indicators of consumption as explanatory variables to construct a fitting and prediction model for the monthly total retail sales of consumer goods. This essay used data from January 2005 to June 2018 for empirical study. The result is good with 1.99% fitting MAPE and 2.39% prediction MAPE. This method is a checksum supplement to the measurement error of the actual total retail sales of consumer goods. At the same time, the model not only contains physical consumption, but also includes the measurement of service consumption. Compared with the index of the total retail sales of consumer goods, it is more comprehensive and real time to reflect the fluctuation of general consumption level. From a macro perspective, this method is a better reference for supply and demand analysis, economic heat, policy effects analysis, and reflection of people's livelihood. Keywords—total retail sales of consumer goods; consumption; prediction