Scholarlink

<p>Global graduate employability is increasingly a concern as higher education enrollment grows worldwide. Traditional employability assessment approaches often fail to capture the complex and multi-dimensional factors that influence graduate success in the labor market. To address these challenges, this research proposes a Modified Elephant Swarm Water Search-driven Deep Deterministic Policy Gradient (MESWS-DDPG) model. The MESWS optimization mechanism is used to tune network parameters and improve convergence efficiency, while the DDPG reinforcement learning framework enables continuous decision-making for dynamically matching graduate profiles with evolving job market requirements. To capture these dimensions, a College Graduates Employment Competitiveness Dataset containing 2,000 records across multiple disciplines is used for training and evaluation with 80% of training and 20% of testing, with data preprocessing including handling missing values and min–max normalization. Additional preprocessing steps include noise reduction, data consistency verification, and feature extraction using Independent Component Analysis (ICA). The model is implemented using Python 3.10 with TensorFlow 2.12 and PyTorch 2.1 frameworks. Experimental results demonstrate that the MESWS-DDPG model outperforms traditional methods and achieves precision (92.54%), recall (94.55%), F-measure (95.87%), and accuracy (93.26%). These results highlight the potential of reinforcement learning–driven optimization frameworks for improving graduate employment competitiveness.</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

An AI-driven reinforcement learning framework for graduate employability prediction and career decision support

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月16日

查看全文
  • Feng Qi,
  • Maixia Liu

摘要

Global graduate employability is increasingly a concern as higher education enrollment grows worldwide. Traditional employability assessment approaches often fail to capture the complex and multi-dimensional factors that influence graduate success in the labor market. To address these challenges, this research proposes a Modified Elephant Swarm Water Search-driven Deep Deterministic Policy Gradient (MESWS-DDPG) model. The MESWS optimization mechanism is used to tune network parameters and improve convergence efficiency, while the DDPG reinforcement learning framework enables continuous decision-making for dynamically matching graduate profiles with evolving job market requirements. To capture these dimensions, a College Graduates Employment Competitiveness Dataset containing 2,000 records across multiple disciplines is used for training and evaluation with 80% of training and 20% of testing, with data preprocessing including handling missing values and min–max normalization. Additional preprocessing steps include noise reduction, data consistency verification, and feature extraction using Independent Component Analysis (ICA). The model is implemented using Python 3.10 with TensorFlow 2.12 and PyTorch 2.1 frameworks. Experimental results demonstrate that the MESWS-DDPG model outperforms traditional methods and achieves precision (92.54%), recall (94.55%), F-measure (95.87%), and accuracy (93.26%). These results highlight the potential of reinforcement learning–driven optimization frameworks for improving graduate employment competitiveness.