Identifying Key Linguistic Features for Emotion Recognition in Text
Giselle Tweeboom
Seisen International School, Tokyo, Japan
Publication date: September 7, 2025
Seisen International School, Tokyo, Japan
Publication date: September 7, 2025
DOI: http://doi.org/10.34614/JIYRC2025I26
ABSTRACT
With the rise of technology and the increasing use of online platforms, the need for accurate emotion recognition in text has become increasingly important. While many emotion detection models focus on identifying the "best" overall approach for recognizing emotions in text, this research aims to determine which specific linguistic features most significantly contribute to this recognition. To explore this, I utilized the ISEAR dataset, which contains 7,666 text samples on various machine-learning models. These models were evaluated for accuracy to identify the best-performing ones, which were then analyzed to determine the contribution of different linguistic categories - nouns, verbs, adverbs, and adjectives - to emotion recognition. The results revealed that adjectives play the most significant role, as they had the highest percentage contribution compared to the other linguistic categories.
With the rise of technology and the increasing use of online platforms, the need for accurate emotion recognition in text has become increasingly important. While many emotion detection models focus on identifying the "best" overall approach for recognizing emotions in text, this research aims to determine which specific linguistic features most significantly contribute to this recognition. To explore this, I utilized the ISEAR dataset, which contains 7,666 text samples on various machine-learning models. These models were evaluated for accuracy to identify the best-performing ones, which were then analyzed to determine the contribution of different linguistic categories - nouns, verbs, adverbs, and adjectives - to emotion recognition. The results revealed that adjectives play the most significant role, as they had the highest percentage contribution compared to the other linguistic categories.
