Knowledge graph construction and talent competency prediction for human resource management

Abstract

Job matching and talent recommendation are essential yet challenging tasks in human resource management. Traditional methods, such as rule-based matching and collaborative filtering, often struggle with issues like data sparsity, cold-start problems, and the dynamic nature of user preferences, limiting their effectiveness in real-world applications. To address these challenges, we propose a hybrid model that integrates Graph Convolutional Networks (GCN), Reinforcement Learning (RL), and Deep Collaborative Filtering (DCF). The GCN module captures complex multi-relational structures between jobs and candidates, the RL module dynamically optimizes recommendation strategies based on feedback, and the DCF module enhances personalized recommendation capabilities. Experimental results demonstrate that the proposed model outperforms traditional methods in key metrics such as Precision@10, Recall@10, NDCG@10, and CTR, while achieving broader coverage. This research provides a novel and effective solution for improving job matching and talent recommendation, offering practical significance for applications in human resource management.