Explainable person-job recommendations: challenges, approaches, and comparative analysis

As person-job recommendation systems (PJRS) increasingly mediate hiring decisions, concerns over their "black box" opacity have sparked demand for explainable AI (XAI) solutions.

This systematic review examines 85 studies on explainable PJRS methods published between 2019 and August 2025, selected from 150 screened articles across Google Scholar, Web of Science, and CNKI, following PRISMA 2020 guidelines.

Guided by a PICOS-formulated review question, we categorize explainability techniques into three layers-data (e.g., feature attribution, causal diagrams), model (e.g., attention mechanisms, knowledge graphs), and output (e.g., SHAP, counterfactuals)-and summarize their objectives, trade-offs, and practical applications. We further synthesize these into an integrated end-to-end framework that addresses opacity across layers and supports traceable recommendations. Quantitative benchmarking of six representative methods (e.g., LIME, attention-based, KG-GNN) reveals performance-explainability trade-offs, with counterfactual approaches achieving the highest Explainability-Performance (E‑P) score (0.95).

This review provides a taxonomy, cross-layer framework, and comparative evidence to inform the design of transparent and trustworthy PJRS systems. Future directions include multimodal causal inference, feedback-driven adaptation, and efficient explainability tools.