Scholarlink

This study delves into the critical realm of employee satisfaction, a subject of paramount importance to corporate managers and leaders. The research proposes an innovative methodology employing natural language processing (NLP) to assess employee sentiment towards an organization through the effective analysis of feedback. The approach incorporates the AdaBoost classifier to discern sentiment encoded utilizing the term frequency-inverse document frequency (TF-IDF) strategy. To enhance classification precision, metaheuristic optimizers are implemented. Moreover, a tailored variant of the SCA metaheuristics has been created to address the weaknesses of the elementary version, aligning it with the specific requirements of this study. The meticulously crafted models demonstrate remarkable performance, achieving a classification accuracy surpassing 88% when verified on a publicly accessible real-world dataset. This research not only contributes a practical and effective approach to gauging employee satisfaction but also showcases the efficacy of combining NLP, machine learning classifiers, and metaheuristic optimization techniques for insightful organizational sentiment analysis.

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

Using Text Mining to Identify Employee Dissatisfaction Optimized by Modified Metaheuristics

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Vule Mizdrakovic,
  • Luka Jovanovic,
  • Miodrag Zivkovic,
  • Milos Antonijevic,
  • Jelena Kaljevic,
  • Nebojsa Bacanin

摘要

This study delves into the critical realm of employee satisfaction, a subject of paramount importance to corporate managers and leaders. The research proposes an innovative methodology employing natural language processing (NLP) to assess employee sentiment towards an organization through the effective analysis of feedback. The approach incorporates the AdaBoost classifier to discern sentiment encoded utilizing the term frequency-inverse document frequency (TF-IDF) strategy. To enhance classification precision, metaheuristic optimizers are implemented. Moreover, a tailored variant of the SCA metaheuristics has been created to address the weaknesses of the elementary version, aligning it with the specific requirements of this study. The meticulously crafted models demonstrate remarkable performance, achieving a classification accuracy surpassing 88% when verified on a publicly accessible real-world dataset. This research not only contributes a practical and effective approach to gauging employee satisfaction but also showcases the efficacy of combining NLP, machine learning classifiers, and metaheuristic optimization techniques for insightful organizational sentiment analysis.