Author(s)

Yongyue Pan, Haoyu Chen, Jiaqi Liu

Corresponding Author

Yongyue Pan

Abstract

This study proposes a hybrid forecasting model com¬bining the Autoregressive Integrated Moving Average (ARIMA) method with the Particle Swarm Optimization (PSO) algorithm to predict China’s pet food production and export trends. By optimizing ARIMA’s parameters using PSO, the hybrid model enhances accuracy and addresses the limitations of traditional linear models. The model utilizes historical data from 2018 to 2023 and provides forecasts for 2024, 2025, and 2026. The results predict significant growth in production, reaching 305.02 million tons, 431.85 million tons, and 597.20 million tons, respectively, re¬flecting the expansion of the domestic market driven by increased pet ownership and demand for premium products. In contrast, export forecasts show a gradual rise to 372.95 billion yuan, 373.73 billion yuan, and 385.75 billion yuan over the same period, indicating stability amidst global trade challenges. These findings provide actionable insights for stakeholders and policymakers, supporting strategic planning and sustainable growth in China’s pet food industry.

Keywords

ARIMA, Particle Swarm Optimization, Pet Food Industry, Time Series Forecasting