Sentiment Analysis for MOOC Course Reviews

Currently, increasing users are using MOOC platforms to choose courses and leave text comments with emotional overtones. The traditional words vector representation method uses static method to extract text information, which ignores the text position information. The traditional convolutional neural network fails to make full use of the semantic features and association information of the text between channels, which will cause inaccurate text sentiment classification. In order to solve the problems, a text classification model based on Albert and Capsule Network and attention mechanism is proposed. The model was verified on the MOOC comment data set, and compared with the traditional user comment sentiment analysis model. The results show the accuracy of the model was improved to a certain extent.