A Dictionary-Based Concept Extraction Method for Chinese Course Knowledge

Chinese Course Concept Extraction holds significant importance in the construction process of knowledge graph in the field of education in China. It aims to extract the concept set of corresponding courses from unstructured texts such as textbooks and course outlines. One of the existing methods for course concept extraction is to encode only character information. However, compared with English, Chinese course concept extraction cannot be separated from contextual language. To tackle this issue, we develop a novel approach named Dictionary-based Chinese Concept Extraction Model, which introduces the word information of the course concept and the professional vocabulary of the third-party database to enrich the representation meaning of character vector. Specifically, first, we construct the course concept dictionary through third-party database such as Baidupedia. Second, each character is matched with word information in the dictionary, which is applied the corresponding weight. Third, the input sequences, represented by character vectors that contain word information, are passed through a single layer of bidirectional Long Short-Term Memory (LSTM) for sequence modeling. Finally, we applies a Conditional Random Field (CRF) layer to infer labels for the entire character sequence. Our proposed method was evaluated on a private dataset, and the results demonstrate its superiority over state-of-the-art methods, through extensive experimentation.