Sentiment Analysis of Student Evaluation of Teaching Based on Bi-LSTM Algorithm
Hongli Yuan, Alexander Arcenio Hernandez · 2023
As the basis of evaluating the teaching quality, the data of students' teaching evaluation play an important role in teaching management and improving teaching quality in universities. Nevertheless, it is challenging to conduct statistical analysis of a substantial number of instructional evaluation texts using manual or conventional technologies. How to deal with and make use of students' teaching evaluation texts efficiently has become an urgent problem to be solved. Based on Bi-directional Long Short-Term Memory (Bi-LSTM), this paper takes students' teaching evaluation texts dataset as the research object to analyze the emotional tendency of students. From the evaluation results, compared with Simple Recursive Neural Network (Simple RNN) and single Long Short-Term Memory (LSTM), the sentiment detection model based on Bi-LSTM has higher accuracy in students' teaching evaluation texts sentiment analysis, the verification accuracy can reach 90.77%, and the testing accuracy can reach 84.33%.