Scholarlink

<p>The escalating strain of depression and its profound consequences on individual and professional well-being underscore the mounting urgency for swift and reliable diagnostic interventions. To address this critical necessity, two pioneering and intricate architectures are proposed for identifying depressive texts in a benchmark textual dataset, leveraging the convergence of machine learning, ensemble learning, and a hybrid swarm optimization, namely SOMRO and Chaos-integrated SOMRO(CI-SOMRO). The proposed SOMRO model endeavors to utilize <b>s</b>eahorse <b>o</b>ptimization’s (SO) proficient exploitative mechanisms and <b>m</b>anta <b>r</b>ay <b>o</b>ptimization’s (MRO) potent exploratory capabilities, harnessing their collective heterogeneity to equilibrate global search dynamics with meticulous local optimization. CI-SOMRO model substantially elevates the performance of SOMRO model by replacing conventional random numbers with dynamic chaotic variants. This infusion of chaos theory empowered the potential of exploitation and exploration processes ultimately fostering the systematic process for finding the best optimal solutions. The comprehensive performance gains of the proposed models are validated by direct comparative assessments against both standalone algorithmic approaches and existing state-of-the-art paradigms. Experimental investigations confirm that the XGBoost ensemble, calibrated with the novel CI-SOMRO model, achieves notable efficacy with an accuracy of 96.29% and F1-score of 95.4%. An elucidation at feature level utilizing the eXplainable Artificial Intelligence (XAI) model LIME (Local Interpretable Model-Agnostic Explanations) is also furnished to support decision-making protocols of clinical psychologists. This research highlights the strong potential of hybrid optimization coupled with machine learning methodologies in promoting reliable and expeditious depression diagnosis, thereby facilitating the refinement of mental health intervention strategies.</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

Unleashing Algorithmic Potential of Chaos-Integrated SOMRO for Enhanced Depression Detection of Textual Data: Optimizing Machine Learning with Hybrid Swarm Intelligence

  • 期刊论文

  • 发表日期：

    2026年6月17日

查看全文
  • Shruty Ahuja,
  • Arunima Jaiswal,
  • Kanchan Sharma

摘要

The escalating strain of depression and its profound consequences on individual and professional well-being underscore the mounting urgency for swift and reliable diagnostic interventions. To address this critical necessity, two pioneering and intricate architectures are proposed for identifying depressive texts in a benchmark textual dataset, leveraging the convergence of machine learning, ensemble learning, and a hybrid swarm optimization, namely SOMRO and Chaos-integrated SOMRO(CI-SOMRO). The proposed SOMRO model endeavors to utilize seahorse optimization’s (SO) proficient exploitative mechanisms and manta ray optimization’s (MRO) potent exploratory capabilities, harnessing their collective heterogeneity to equilibrate global search dynamics with meticulous local optimization. CI-SOMRO model substantially elevates the performance of SOMRO model by replacing conventional random numbers with dynamic chaotic variants. This infusion of chaos theory empowered the potential of exploitation and exploration processes ultimately fostering the systematic process for finding the best optimal solutions. The comprehensive performance gains of the proposed models are validated by direct comparative assessments against both standalone algorithmic approaches and existing state-of-the-art paradigms. Experimental investigations confirm that the XGBoost ensemble, calibrated with the novel CI-SOMRO model, achieves notable efficacy with an accuracy of 96.29% and F1-score of 95.4%. An elucidation at feature level utilizing the eXplainable Artificial Intelligence (XAI) model LIME (Local Interpretable Model-Agnostic Explanations) is also furnished to support decision-making protocols of clinical psychologists. This research highlights the strong potential of hybrid optimization coupled with machine learning methodologies in promoting reliable and expeditious depression diagnosis, thereby facilitating the refinement of mental health intervention strategies.