GRAPH NEURAL NETWORKS FOR COMPETENCY MAPPING: A DEEP LEARNING APPROACH TO EMPLOYEE SKILL DEVELOPMENT

Employee skill development is an important aspect of workforce optimization in Human Resource Management.Classical competency mapping models do not account for complex skill relationships and will often recommendsuboptimal training. To overcome this, we introduce a Graph Convolutional Network based framework forcompetency mapping that accurately captures job-related skill interdependencies. With the Related Job Skillsdataset, we build a skill graph whose nodes are skills and edges are their relations based on industry trends andchanging job requirements. The model learns skilled representations, allowing correct prediction of upskillingtrajectories. The presented framework supports more informed HR decision-making by determining best-fittingskill development opportunities, enhancing flexibility, and supporting career advancement. Experimentalfindings illustrate that our model attains accuracy of 97.3%, precision of 95.1%, recall of 96.4%, and F1-scoreof 95.7%, surpassing the conventional methods. Moreover, our framework obtains MRR of 96.8% and NDCGof 97.2%, which guarantees the quality of the recommended skills. Computational efficiency of the model witha mean training time of 0.9 seconds per epoch allows it to be scalable with big HR data and real-time systems.This work presents a strong data-driven HRM solution with guaranteed strategic workforce planning andongoing employee development. In future, effort will be dedicated to graph architecture optimization andenlarging datasets to achieve better generalization and adoption in real-world applications.