Scholarlink

<p>Employment competitiveness analysis is complex and ambiguous due to human characteristics and labor market conditions. Existing evaluation systems use static scoring or deterministic approaches and cannot capture causal links and dynamic interactions between indicators. This study presents a Multi-Dimensional Employment Competitiveness Evaluation Framework (MDECEF-BN) to address these concerns. The framework uses a Bayesian Network to assess the interdependencies of school background, professional skills, job experience, adaptability, and labor-market circumstances in a hierarchical structure. Expert knowledge informs the network structure, while survey data and employment records feed the conditional probability tables. Probabilistic competitiveness assessments provide uncertainty propagation, scenario reasoning, and sensitivity analysis. The study proposes a new multi-dimensional model of systematic analysis of employment competitiveness, uses Bayesian Network modeling to understand complex links between employment indicators, and provides insights and forecasting ability in talent management, labor planning, and employment advice systems. The proposed method compares the framework to standard weighted scoring, the Analytic Hierarchy Process (AHP), and logistic regression utilizing F1-score and MAE, interpretability consistency, and performance. Findings indicate that ECEF-BN has an accuracy of 87.6, which is more than weighted scoring (74.2%), AHP (78.9%), and logistic regression (82.1), and decreases MAE by 18.25% and increases F1-score by 94% on average. The most significant influences are work experience and professional skills, which together account for more than 45% of the variability in competitiveness. These results demonstrate that ECEF-BN is more accurate, consistent, and interpretable; is evidence-based; and can be generalized in subsequent research to dynamic and/or longitudinal employment data.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A multi-dimensional evaluation framework for employment competitiveness based on Bayesian networks

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月7日

查看全文
  • Kai Liu

摘要

Employment competitiveness analysis is complex and ambiguous due to human characteristics and labor market conditions. Existing evaluation systems use static scoring or deterministic approaches and cannot capture causal links and dynamic interactions between indicators. This study presents a Multi-Dimensional Employment Competitiveness Evaluation Framework (MDECEF-BN) to address these concerns. The framework uses a Bayesian Network to assess the interdependencies of school background, professional skills, job experience, adaptability, and labor-market circumstances in a hierarchical structure. Expert knowledge informs the network structure, while survey data and employment records feed the conditional probability tables. Probabilistic competitiveness assessments provide uncertainty propagation, scenario reasoning, and sensitivity analysis. The study proposes a new multi-dimensional model of systematic analysis of employment competitiveness, uses Bayesian Network modeling to understand complex links between employment indicators, and provides insights and forecasting ability in talent management, labor planning, and employment advice systems. The proposed method compares the framework to standard weighted scoring, the Analytic Hierarchy Process (AHP), and logistic regression utilizing F1-score and MAE, interpretability consistency, and performance. Findings indicate that ECEF-BN has an accuracy of 87.6, which is more than weighted scoring (74.2%), AHP (78.9%), and logistic regression (82.1), and decreases MAE by 18.25% and increases F1-score by 94% on average. The most significant influences are work experience and professional skills, which together account for more than 45% of the variability in competitiveness. These results demonstrate that ECEF-BN is more accurate, consistent, and interpretable; is evidence-based; and can be generalized in subsequent research to dynamic and/or longitudinal employment data.