A Study of Knowledge Discovery and Pattern Recognition Based on Large-Scale Sentiment Data in Online Education for College Students

Abstract

With the development of the internet, online education platforms have become the main means of acquiring knowledge for people. To improve the efficiency of user analysis, this paper proposes a framework for user sentiment recognition based on course review information of online education platforms. Firstly, according to the characteristics of the review data of online platforms, the authors use crawlers and open-source word separation tools to complete data collection and collation; secondly, the authors use TextCNN (text convolutional neural networks) and BILSTM (bi-long-short-term-memory) to combine to build a feature layer fusion text sentiment classification model. The results show that the average precision of the model built is 95.1% for the three sentiment classifications. Finally, a test of the reviews of new courses on online education platforms is conducted. The sentiment recognition rate of all three types of new courses exceeds 90%. Therefore, the proposed model provides new ideas for user analysis.