A prediction model for cross-cultural communication effects of Shu Road culture integrating knowledge graph and attention mechanism

Abstract

In western China, Shu Road culture is a typical example of regional trade, transportation, and cultural fusion. Semantic fragmentation, audience cognitive differences, and a tardy effect evaluation are still present in cross-cultural communication. In this article, provide a prediction model that combines an attention-based feature fusion mechanism with a Shu Road cultural knowledge graph. Create a structured semantic network that includes cultural items, historical locations, route events, story themes, media outlets, and audience response. Estimate the frequency of communication, sharing, cultural familiarity, and popularity of users on this site. The suggested model outperforms traditional machine learning, BiLSTM, Transformer, and graph neural network baselines with an 87.6% prediction accuracy and a macro-F1 of 85.9%, according to experimental study based on multilingual cultural material and simulated platform data. The communication effect is influenced by visual-cultural symbols, historical route semantics, attention analysis, and language adaptation.