Scholarlink

<p>Occupational stress is a complex and multifactorial condition resulting from excessive work demands, role ambiguity, and inadequate coping mechanisms, significantly affecting employee health and organizational performance. The evaluation of occupational stress levels in humans requires the processing of high-dimensional, heterogeneous, and intrinsically inconsistent physiological and psychological datasets, posing significant challenges to the efficiency and generalization capability of conventional machine learning techniques. Swarm Intelligence (SI) algorithms, modeled on the self-organizing collective dynamics observed in natural systems such as bird flocking and ant foraging, offer robust metaheuristic strategies for high-dimensional optimization and feature selection in complex problem domains. Their intrinsic capability to maintain a balance between exploration and exploitation of the search space facilitates the identification of highly discriminative stress-related attributes while effectively reducing redundancy. Furthermore, SI algorithms demonstrate scalability, adaptability to dynamic environments, and competence in capturing complex non-linear interactions, making them well-suited for delivering accurate, robust, and interpretable diagnostic outcomes in occupational stress assessment. This study introduces a novel swarm intelligence-based algorithm, Adaptive Grey Wolf Optimization with Restricted Crow Search Method (AGWO_RCSM), designed for optimal feature extraction in real-world occupational stress assessment scenarios. The performance of the proposed approach has been rigorously evaluated using multiple evaluation metrics on the SWELL-KW dataset to assess its effectiveness and robustness. Furthermore, three publicly available benchmark datasets, namely Arrhythmia, Parkinson’s Disease, and PIMA Diabetes, have been also exhaustively analyzed using identical experimental settings. Although the benchmark datasets originate from different application domains, they encompass diverse classification characteristics that facilitate a rigorous assessment of the robustness, adaptability, and transferability of the proposed optimization framework across heterogeneous data distributions.</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 swarm intelligence driven model for occupational stress diagnosis: design, optimization and analysis

  • 期刊论文

  • 发表日期：

    2026年8月3日

查看全文
  • Samriti Sharma,
  • Gurvinder Singh

摘要

Occupational stress is a complex and multifactorial condition resulting from excessive work demands, role ambiguity, and inadequate coping mechanisms, significantly affecting employee health and organizational performance. The evaluation of occupational stress levels in humans requires the processing of high-dimensional, heterogeneous, and intrinsically inconsistent physiological and psychological datasets, posing significant challenges to the efficiency and generalization capability of conventional machine learning techniques. Swarm Intelligence (SI) algorithms, modeled on the self-organizing collective dynamics observed in natural systems such as bird flocking and ant foraging, offer robust metaheuristic strategies for high-dimensional optimization and feature selection in complex problem domains. Their intrinsic capability to maintain a balance between exploration and exploitation of the search space facilitates the identification of highly discriminative stress-related attributes while effectively reducing redundancy. Furthermore, SI algorithms demonstrate scalability, adaptability to dynamic environments, and competence in capturing complex non-linear interactions, making them well-suited for delivering accurate, robust, and interpretable diagnostic outcomes in occupational stress assessment. This study introduces a novel swarm intelligence-based algorithm, Adaptive Grey Wolf Optimization with Restricted Crow Search Method (AGWO_RCSM), designed for optimal feature extraction in real-world occupational stress assessment scenarios. The performance of the proposed approach has been rigorously evaluated using multiple evaluation metrics on the SWELL-KW dataset to assess its effectiveness and robustness. Furthermore, three publicly available benchmark datasets, namely Arrhythmia, Parkinson’s Disease, and PIMA Diabetes, have been also exhaustively analyzed using identical experimental settings. Although the benchmark datasets originate from different application domains, they encompass diverse classification characteristics that facilitate a rigorous assessment of the robustness, adaptability, and transferability of the proposed optimization framework across heterogeneous data distributions.