Implementation Of Emotion Analysis Algorithm in Analysing Public Cognition of Industrial Heritage

Abstract

Industrial heritage sites hold deep historical and cultural significance, yet public perceptions of their value vary widely. Visitors often form emotional connections to such sites, offering valuable insights for tourism enhancement and cultural preservation. Traditional methods fall short in effectively capturing public sentiment toward industrial heritage. To address this gap, we propose the Advanced Northern Goshawk-Responsive Generative Adversarial Network (ANG-RGAN) technique, which leverages emotion analysis to assess public sentiment. ANG-RGAN combines Responsive Generative Adversarial Networks with attention mechanisms, improving sentiment classification accuracy. Surveys and interviews conducted at Beijing’s industrial heritage sites provided textual data for analysis. By employing tokenization, normalization, and term frequency-inverse document frequency (TF-IDF), emotional expressions such as curiosity, nostalgia, and melancholy were extracted and classified. Experimental results demonstrate superior performance, with an F1-score of 98.3%, accuracy of 98.9%, precision of 98.2%, and recall of 98.5%. Compared to existing methods, ANG-RGAN offers enhanced security and reliability in emotion-based assessments. This approach provides critical insights into public perceptions, fostering the growth of the tourism industry and the preservation of industrial heritage.