Medical artificial intelligence in scholarly and public perspective: BERTopic-based analysis of topic-sentiment collaborative mining

In order to deeply study the development trend of AI application in the medical field and the responses of different groups of people, based on the literature data of scholars and public opinion data in the field of medical artificial intelligence, this study combines the BERTopic model with SnowNLP for theme-sentiment collaborative mining analysis. Firstly, the topic model is used to analyze the features and differences of the topics in depth, and get the topic modelling results, so as to reveal the overall research structure of the field. Then, the sentiment analysis of each theme was conducted to obtain the sentiment classification under the corresponding theme text, in order to explore the emotional responses of different perspectives on this domain. This study not only considers the theme distribution and sentiment tendency at the same time, but also explores the trend of sentiment evolution and influencing factors through the exploration, with a view to providing useful references for future research. The results of the study show that academics and the public are interested in different topics. Scholars have more positive affective tendencies towards most topics, while the public has relatively lower positive affective tendencies. The public had a large percentage of negative emotions for the same topics compared to academics.