A session-based job recommendation system combining area knowledge and interest graph neural networks

Online job boards become one of the central components of the modern recruitment industry. Existing systems are mainly focused on content analysis of resumes and job descriptions, so they heavily rely on the accuracy of semantic analysis and the coverage of content modeling, in which case they usually suffer from rigidity and the lack of implicit semantic relations. In recent years, session recommendation has attracted the attention of many researchers, as it can judge the user's interest preferences and recommend items based on the user's historical clicks. Most existing session-based recommendation systems are insufficient to obtain accurate user vectors in sessions and neglect complex transitions of items. We propose a novel method, Area Knowledge and Interest Graph Neural Networks(AIGNN). We add job area knowledge to job session recommendations, in which session sequences are modeled as graph-structured data, then GNN can capture complex transitions of items. Moreover, the attention mechanism is introduced to represent the user's interest. Experiments on real-world data set prove that the model we proposed better than other algorithms.