Research on pet market forecasting problem based on multiple linear regression and ARIMA model

The research on pet market forecasting is to use appropriate models and algorithms to predict future information, which is of great significance for the pet industry to formulate development strategy and improve development quality. This paper first assumes that pet market is related to GDP, consumption level of urban residents, per capita disposable income of residents, total retail sales of social consumer goods, added value from primary industry to tertiary industry, number of births, population aging, and dozens of influencing factors. After collecting and sorting relevant data of China from 1990 to 2023, a multiple linear regression prediction model was constructed, and appropriate independent variables were determined through correlation analysis, F-test and t-test. After determining the independent variable factors, the time series ARIMA model is used to predict the independent variable data, and then the multiple linear regression model is introduced to predict the pet market data. Through the calculation of China's pet market forecast in the next five years, the forecast effect is good. The forecast method in this paper is of great significance to the development of pet-related industries and plays a great role in improving the development quality of pet market.