"基于BERT的淮安旅游景点评论情感分析研究"

Graduation thesis is based on bert's text sentiment analysis and public opinion analysis template. Huaiyin Institute of Technology graduation thesis instructions (thesis) May 2021 graduation design explanation (thesis) Chinese abstract User comments on tourism websites can truly reflect tourists' feelings about the service and travel experience of tourist attractions after their travels. This topic extracts the opinions of tourists on tourist attractions to understand the core concerns of tourists on tourist attractions, uses sentiment analysis to understand the preferences of tourists for each tourist attraction, uses clustering classification capabilities to systematically summarize, and finally displays the attractions according to the preferences of each core concern in a visual form. Among them, sentiment analysis is implemented by constructing a BERT model, and the accuracy of the training set and test set reaches 0.96. Testing on a self-built Huai'an tourism comment dataset, experimental tests show that the sentiment analysis model is successfully constructed, can predict the sentiment orientation of comments with high accuracy, and visually analyze and present the sentiment orientation of each core concern of Huai'an tourist attractions, thereby intuitively showing the level of tourists' love for each tourist attraction in Huai'an, achieving the desired goal. Keywords Sentiment analysis; BERT; text classification; deep learning Graduation design explanation (thesis) Foreign abstract Title Sentiment analysis of social media comments for Huai'an Abstract Regarding to: "。。".