Abstract: The rapid development of artificial intelligence has promoted the intellectualization in the field of education, and research on the construction of subject knowledge system by using knowledge graphs has emerged. The knowledge system constructed by knowledge graphs can model the correlation among knowledge points and form the course knowledge context, which is helpful for learners to memorize and understand knowledge in a deep level. However, the teaching resources of subject courses are diverse and scattered, and the existing automatic construction methods often use a single data resource with low resource utilization rate, which is difficult to provide guidance for the construction of professional knowledge systems. Moreover, low degree of automation and high cost need different modules to process the unstructured texts and generate the graphs, so reproducibility and portability are weak. Therefore, an optimization process of subject knowledge graph construction is proposed based on the large language model, which highly integrates the advantages of the large language model and the logical correlation of knowledge in various subjects. Specifically, taking the major of communication as an example, the powerful function of a large language model such as ChatGPT is used to design a knowledge system through subject knowledge analysis by using a variety of knowledge sources, so as to automatically extract knowledge entities and knowledge relations based on teaching resources to form a final subject knowledge system, which can also be introduced to other subject fields. 

Key words:  large language model, knowledge graph, intelligent education, ChatGPT, automated graph construction