Sentiment Distribution of Topic Discussion in Online English Learning

Abstract

Online English teaching resources have recently surged, highlighting the exigency for efficient organization and categorization. This manuscript introduces an innovative strategy to classify university-level English teaching resources, employing a sophisticated density clustering algorithm. Initially, student discourse was mined within a teaching platform comment section, and in-depth textual analysis was conducted. Subsequently, the term frequency-inverse document frequency (TF–IDF) feature extraction algorithm was enhanced, while emotive attributes were seamlessly integrated into the textual manifestation layer during the classification procedure. This enabled the distribution of topics and emotions to be acquired for each comment, facilitating subsequent analyses of emotion feature extraction and model training. An improved weight calculation was designed based on TF–IDF to evaluate the importance of feature items for each corpus file. The simulation results demonstrate the proposed scheme's effectiveness. The algorithm facilitates faster scholarly access to educational resource information and effectively classifies data for high research adaptability.