Prediction of Student Attitude towards Blended Learning Based on Sentiment Analysis

Blended learning, serving as a popular and effective teaching means, attracts increasing attentions for both of the educators and researchers. To achieve better learning performance, the blended learning makes full use of the advantages of both e-learning and face-to-face instruction. Traditional educational researches and excises of the blended learning tend to pay more attentions on the objective improvements of student learning achievements. Nevertheless, the subjective attitude and feedback from the student perspective are prone to being neglected. In the paper, we make the effort towards the attitude prediction of student in the blended learning. Especially, the machined learning based sentiment analysis is introduced to improve the prediction performance. To this end, both of the categorical and textual features are properly extracted. In particular, the word embedding technique is adopted to transform the textual data into context-sensitive vector representation. Several typical classifiers are examined to obtain corresponding superior prediction outcomes. The evaluation results demonstrate that the student attitude for a targeted blended learning class can be precisely predicted. In addition, the sentiment analysis factors for the prediction are also carefully discussed.