Research On Fine-grained Sentiment Analysis Based On Students' Evaluation Of Teaching Text

With the continuous expansion of the scale of higher education in China,the society pays more and more attention to the quality of teaching in colleges and universities.Students' evaluation of teaching is an extremely important part of current teaching quality evaluation.With the continuous accumulation of students' evaluation of teaching data,the mining and utilization of students' evaluation of teaching data has gradually become the focus of research.The text of students' evaluation of teaching contains students' specific attitudes towards various aspects of teacher teaching performance.If it is deeply explored and utilized,it will be beneficial to find out the specific deficiencies of teachers' teaching process so as to carry out targeted improvement and improve the teaching quality.The purpose of this paper is to conduct a fine-grained sentiment analysis of the texts of students' evaluation of teaching,to mine and feed back the teaching evaluation dimensions and their corresponding emotional tendencies in students' evaluation of teaching texts.The main work contents are as follows:(1)Defined a student-centered teaching evaluation dimension label system by analyzing a large number of real students' teaching evaluation data;(2)Proposed a method for identifying evaluation dimensions based on feature word matching,by constructing feature extraction rules of dependency syntax analysis and designing a matching evaluation strategy for teaching dimensions based on seed vocabulary,thereby identifying multiple teaching evaluation dimensions in the teaching evaluation text;(3)Proposed a staged sentiment analysis method based on the combination of teaching evaluation emotion dictionary and based on DEA Bi-LSTM network model to realize the emotion judgment of different teaching evaluation dimensions;(4)Designed an overall plan for students' evaluation of teaching feedback analysis based on fine-grained opinion mining,and proposed a viewpoint integration method based on sentiment tags and a feedback analysis method based on attention and sentiment scores to further improve the application value of this research.Finally,experiments show that the teaching evaluation dimension recognition method and the sentiment analysis method for teaching evaluation dimension proposed in this paper caneffectively identify multiple teaching evaluation dimensions and their corresponding emotional tendencies in student evaluation texts,and the F1-score of the combination of the two modules reaches 86.16% and 85.49% respectively.This paper demonstrates the effect of the feedback analysis plan for students' evaluation of teaching based on the fine-grained sentiment analysis in the form of case analysis.It can be seen that the plan can analyze the teaching quality of teachers from multiple dimensions,and has good practical significance and application value.