TOPSIS and VAR-Based Predictive Modeling for the Pet Sector: Integrated Evaluation and Forecasting

Abstract

Investigating the dynamic evolution and consumer-driven demands of the pet sector is pivotal for mapping industry trajectories, uncovering latent market opportunities, aligning with user expectations, refining product portfolios, designing adaptive marketing tactics, and accelerating sectoral innovation. This not only helps businesses and investors make scientific decisions but also fosters the sustained and sound development of the pet sector. Therefore, this paper proposes a comprehensive evaluation and forecast model for the trends in development and market demand of the pet sector based on the TOPSIS and VAR algorithms. Firstly, this paper selects five appropriate indicators by pet category and establishes a comprehensive evaluation model for the Chinese cat and dog market using TOPSIS and the entropy weight method. It then conducts k-means++ clustering based on the development scores of the cat and dog market in recent years. Secondly, using the Spearman correlation method, indicators with a correlation coefficient greater than 0.9 are identified as strongly correlated factors for the development of China's pet industry. Based on ARIMA, the values of these strongly correlated factors over the next three years are predicted, and the development situation of China's pet industry market in the coming three years is calculated, with the factor showing the largest increase reaching 35%. Then, the VAR model is used to apply positive shocks to tariff policy factors, monetary policy factors, and international trade factors, observing the quantitative changes in the impacted variable, which is China's pet food industry, to formulate feasible strategies to promote the growth of China's pet food sector. Finally, this paper discusses and analyzes the established model, comprehensively evaluating its advantages and disadvantages.