“…He predicted social logistics volume by constructing a prediction model that combined BP neural network and principal component regression analysis [14]. At the level of fresh agricultural product consumption, Wang S selected 14 indicators affecting fresh agricultural product consumption from five aspects: regional development level, market supply and demand factors, industry structure level, location advantage factors, and logistics industry factors, and then constructed a combined prediction model based on SVR [15]. Wang X predicted fresh agricultural product consumption from five aspects: agricultural product supply, cold chain level, socio-economic indicators, logistics demand scale, and human development angle, and optimized neural network using genetic algorithm [16].…”