Recommendation of Micro Teaching Video Resources Based on Topic Mining and Sentiment Analysis

Abstract

Video learning resources are preferred by many students, owing to their intuitiveness and attractiveness. It is of practical significance to study the recommendation methods of video learning resources. Most of the existing research methods treat the scoring matrix as the main element, failing to consider video contents and learner interests. As a result, few of them can realize precise recommendation of videos. To solve the problem, this paper explores the recommendation of micro teaching video resources based on topic mining and sentiment analysis. Firstly, the dialog text features of English dialog videos and learner interest features were mined based on the deep word vector, and a topic mining model was established to achieve similarity-based resource recommendation. Next, the micro teaching videos with text information were subjected to sentiment analysis, improving the pushing accuracy of micro teaching videos. Finally, the scientific nature of our algorithm was demonstrated through experiments.