OBE Oriented Constructing Method of Educational Knowledge Graph for Automatic Control Discipline

With the rapid development of education informatization and the exponential growth of education big data, Educational Knowledge Graph (EKG) has received significant attention in the field of smart education. However, current research of constructing Chinese EKG still faces enormous challenges, such as incomplete education ontology, low recognition performance of educational entities and relations, and the scarce dataset. This article proposes an OBE (Outcome-Based Education) oriented constructing method of Chinese EKG for automatic control discipline. We first design an OBE based multilevel education ontology for EKG to guide the knowledge modeling of automatic control discipline. And then, we propose a DWBERT-BiLSTM-CRF model to recognize more specific educational entities, and combines BERT-wwm with BiGRU and attention mechanism to efficiently extract the entity relations. In addition, we construct a Chinese dataset for identifying educational entity and its relations of automatic control discipline (ACEER). We evaluate the effectiveness and generalizations of our EKG model by extensive experiments on our ACEER dataset and four benchmark datasets, which demonstrates that our proposed models can achieve outstanding performance on all datasets.