Research On User Satisfaction Of Digital Services In Public Libraries Based On Aspect Level Emotional Analysis

There are more and more comments and reviews on the digitization of libraries by users on the Internet.These comments are useful to improve user satisfaction of library digital reading services.Manually analyzing the large number of reviews is practically impossible.Therefore,to solve this problem,we employ the theoretical approach of deep learning to mine the overall sentiment or opinion polarity from reviews.This paper adopts sentiment analysis and text mining technology to analyze library digitization.Most of the previous work has been done in document or sentence level sentiment analysis.Whereas library review data sentiment scores are assigned to aspects based on the words used.This paper mainly studies opinion mining at various levels of digital library reviews.This paper utilizes word2vec’s Chinese keyword extraction technique to construct an aspect word thesaurus,employing sentence extraction technology to extract aspect word data.Additionally,PMI+SKEP model is utilized to label emotions automatically.To compare the attention mechanism algorithm with multiple groups,the most efficient deep learning algorithm is determined to predict user satisfaction and draw corresponding conclusions.According to the obtained satisfaction results,scientific suggestions are provided for the digital construction of the library.