Knowledge Points Extraction of Junior High School English Exercises Based on SVM Method

Abstract

In the process of learning English, students need to do a lot of exercises to improve English performance. The knowledge points of exercises are important to students, yet how to extract the knowledge points from exercises automatically is difficult, which is the foundation of the knowledge graph construction for students learning. In this paper, we use SVM to realize the knowledge points extraction of junior high school English exercises. Firstly, this paper obtains amounts of question data through analyzing electronic documents, and uses NLP tools to segment, POS tagging and named entity recognition. Secondly, we extract the knowledge points based on SVM model, which involves building multi-class feature vectors and constructing a hierarchical classification for question data. Finally, the experimental results prove the method is effective.