Scholarlink

<p>Emotions play a crucial role in human communication, influencing interactions and decision-making processes. Integrating emotional awareness into machines is increasingly important, as it enables more natural and effective interactions between humans and machines. The proposed work focuses on emotion detection using the ISEAR dataset. During preprocessing, word-intensity lexicon is utilized to extract the features, which were then categorized into positive and negative groups. A Class Topper Optimization (CTO) based Gated Recurrent Unit (GRU) model is proposed to classify the sentences into two groups, effectively framing the task as a binary classification problem. The performance of the proposed model is evaluated against other models using various metrics and resulted in 98.87% accuracy, 98.51% precision, 100% recall, 95.58% specificity and 99.25% F1 score, mean squared error 2.23%, root mean squared error 15.99% and mean absolute error 8.57% and outperformed some existing models, providing more accurate and reliable classifications. This advancement in emotion detection is significant, as it enhances the ability to understand and respond to human emotions, which is essential for applications in fields such as mental health, human-computer interactions and social media analysis.</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 CTO-based GRU model for identifying emotions from textual data

  • 期刊论文

  • 发表日期：

    2025年2月21日

查看全文
  • Shyam Sunder Jannu Soloman,
  • Behilo Seb,
  • Nagaraju Baydeti,
  • Dushmanta Kumar Das

摘要

Emotions play a crucial role in human communication, influencing interactions and decision-making processes. Integrating emotional awareness into machines is increasingly important, as it enables more natural and effective interactions between humans and machines. The proposed work focuses on emotion detection using the ISEAR dataset. During preprocessing, word-intensity lexicon is utilized to extract the features, which were then categorized into positive and negative groups. A Class Topper Optimization (CTO) based Gated Recurrent Unit (GRU) model is proposed to classify the sentences into two groups, effectively framing the task as a binary classification problem. The performance of the proposed model is evaluated against other models using various metrics and resulted in 98.87% accuracy, 98.51% precision, 100% recall, 95.58% specificity and 99.25% F1 score, mean squared error 2.23%, root mean squared error 15.99% and mean absolute error 8.57% and outperformed some existing models, providing more accurate and reliable classifications. This advancement in emotion detection is significant, as it enhances the ability to understand and respond to human emotions, which is essential for applications in fields such as mental health, human-computer interactions and social media analysis.