A Survey of Methods and Architectures for Attention-Based Sparse Graph Convolutional Neural Network-Based Forecast Model for Career Planning in Human Resource Management

Abstract

The rapid advancement of artificial intelligence and big data analytics has significantly transformed human resource management (HRM), particularly in career planning and workforce analytics. Traditional HR systems, which rely on rule-based or statistical methods, often fail to capture the complex relationships among employees, skills, job roles, and organizational structures. With the increasing availability of workforce data from enterprise systems and professional platforms, intelligent predictive models have become essential. Graph-based machine learning approaches, especially Graph Convolutional Neural Networks (GCNs), effectively model such relationships by representing entities as nodes and interactions as edges. These models enable accurate forecasting of career paths and personalized recommendations. However, conventional GCNs treat all connections equally, limiting their effectiveness. Attention mechanisms address this by assigning importance to relevant nodes and relationships, improving prediction accuracy and interpretability. Additionally, sparse graph neural networks handle incomplete and heterogeneous HR data efficiently. Attention-based and heterogeneous graph models further enhance job recommendation systems by capturing complex dependencies between skills, roles, and career trajectories. This review highlights advancements in attention-based sparse GCN architectures for career forecasting, emphasizing their ability to support intelligent talent management, while also identifying challenges and future directions for scalable and adaptive HR analytics systems.