Scholarlink

The study presents a sentiment analysis framework to evaluate employee feedback systematically, providing actionable insights for organizational improvement. The model classifies sentiments into positive and negative categories with an accuracy of 80%, supported by a macro-average F1-Score of 0.75. Positive sentiments exhibited high recall, while negative sentiments showed lower recall, indicating room for enhancement in detecting dissatisfaction. Aspect-level Sentiment analysis revealed varied feedback trends, with the highest satisfaction observed in Aspect 0 and dissatisfaction in Aspect 4 and 8. Visualizations including a bar chart and spider plot highlighted these trends enabling stakeholders to identify strengths and address shortcomings. The framework empowers organizations to analyze employee sentiment, prioritize interventions, and make data-driven decisions. Future improvements could refine sentiment classification and integrate dynamic feedback tracking, ensuring continuous optimization of workplace policies and practices.

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 Unified Framework for Interpretable Sentiment Analysis Using Explainable AI

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • R. V. Dhanalakshmi,
  • M. Shashank,
  • K. Sushmita,
  • K. J. Nithin,
  • S. N. Keerthana,
  • Shivashish Gour

摘要

The study presents a sentiment analysis framework to evaluate employee feedback systematically, providing actionable insights for organizational improvement. The model classifies sentiments into positive and negative categories with an accuracy of 80%, supported by a macro-average F1-Score of 0.75. Positive sentiments exhibited high recall, while negative sentiments showed lower recall, indicating room for enhancement in detecting dissatisfaction. Aspect-level Sentiment analysis revealed varied feedback trends, with the highest satisfaction observed in Aspect 0 and dissatisfaction in Aspect 4 and 8. Visualizations including a bar chart and spider plot highlighted these trends enabling stakeholders to identify strengths and address shortcomings. The framework empowers organizations to analyze employee sentiment, prioritize interventions, and make data-driven decisions. Future improvements could refine sentiment classification and integrate dynamic feedback tracking, ensuring continuous optimization of workplace policies and practices.