Scholarlink

<p>Understanding human emotions is of paramount importance in security, healthcare, and educational fields. Although nonverbal communication—making use of facial expressions, head gesture, and eye contact—comprises most of the interpersonal interaction, the rising technology of AI in the emotional recognition is still in its early days. In this work we examine the Human Emotion Recognition (HER) technology that has software tool for detection and classification of emotions by means of regression, classification {SMO, J48, Multilayer Perceptron} and rule based learning {Apriori} algorithms. This work may help to connect human expertise and computational resources in the service of affective analysis in naturalistic contexts. Experimental results, validated through expert annotation, show that interpretable classification models—particularly J48—produce decision structures that are logically consistent with expert reasoning. Rather than claiming generalizable accuracy, the proposed framework is positioned as an interpretability-driven proof-of-concept that demonstrates how symbolic reasoning and human validation can be integrated into emotion-aware AI systems.</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

From perception to interpretation: a replicable AI framework for emotion recognition based on human validation and Apriori association rules

  • 期刊论文

  • 发表日期：

    2026年4月30日

查看全文
  • Pablo Vicente Torres,
  • Erika Vanessa Cabrera Urbina,
  • Silvio Machuca Vivar,
  • Luis Molina Chalacan,
  • Edgar Freddy Robalino Peña,
  • Susana A. Arias T.

摘要

Understanding human emotions is of paramount importance in security, healthcare, and educational fields. Although nonverbal communication—making use of facial expressions, head gesture, and eye contact—comprises most of the interpersonal interaction, the rising technology of AI in the emotional recognition is still in its early days. In this work we examine the Human Emotion Recognition (HER) technology that has software tool for detection and classification of emotions by means of regression, classification {SMO, J48, Multilayer Perceptron} and rule based learning {Apriori} algorithms. This work may help to connect human expertise and computational resources in the service of affective analysis in naturalistic contexts. Experimental results, validated through expert annotation, show that interpretable classification models—particularly J48—produce decision structures that are logically consistent with expert reasoning. Rather than claiming generalizable accuracy, the proposed framework is positioned as an interpretability-driven proof-of-concept that demonstrates how symbolic reasoning and human validation can be integrated into emotion-aware AI systems.