Regional synergetic effects of ice and snow consumption based on a BERT neural network

This study takes Harbin Ice-Snow World as a key case and develops an integrated framework based on a BERT neural network and regional economic analysis to quantify and analyze regional synergy associated with ice-snow consumption. The framework integrates semantic features with structured indicators to construct a regional synergy intensity model using approximately 12,000 consumer reviews and multidimensional statistical data from 15 regional nodes. Five-fold cross-validation and a three-expert consistency assessment (Fleiss’ Kappa = 0.84) were conducted. The results showed that the model achieved strong performance in predicting synergetic relationships, with an F1-score of 0.87, significantly outperforming baseline models based on advanced large language models such as Llama-3-70B and the ablation variants. Harbin was identified as the “absolute core” of the network, with a weighted centrality of 4.52. Based on quantitative topological indicators, including network density and path strength, the synergetic patterns were classified into the “Heilongjiang Core Agglomeration Zone,” the “Jilin-Liaoning-Hebei Functional Complementarity Zone,” and the “Long-Distance Characteristic Linkage Zone.” This study provides a new perspective for understanding the spatial organization of the ice-snow industry and offers data-driven decision support for formulating cross-regional collaboration policies.