Emotion Recognition based on Text Summarization and Word Features

This paper proposes a novel emotion recognition model that integrates text summarization for global semantics and word-level affective features for temporal dynamics. The model features two dedicated modules: a Semantic Understanding Module (SUM) that deliberately abstracts from temporal order to distill stable, global semantics from noisy, lengthy text via K-means-based summarization and BERT encoding; and an Emotion Perception Module (EPM) that explicitly captures chronological emotion evolution by extracting lexicon-based features from the post sequence and modeling them with a BiGRU. Evaluated on both the GoEmotions and EmoInt datasets, our method achieves superior performance. Ablation studies confirm each module's critical role. The results demonstrate the effectiveness of combining structured semantic comprehension with fine-grained emotion dynamics for accurate recognition in user-generated text.

• Publication year

  2025

• Venue

  2025 5th International Conference on Communication Technology and Information Technology (ICCTIT)

• Publication date

  2025-12-26

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1109/ICCTIT68197.2025.11406421
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Challenges and Opportunities of Text-Based Emotion Detection: A Survey

  2024cited by this paper
• A Survey of Deep Learning-Based Multimodal Emotion Recognition: Speech, Text, and Face

  2023cited by this paper
• Knowledge-based BERT word embedding fine-tuning for emotion recognition

  2023cited by this paper
• Text Emotion Recognition Based on XLNet-BiGRU-Att

  2023cited by this paper
• Textual emotion recognition method based on ALBERT-BiLSTM model and SVM-NB classification

  2023cited by this paper
• Text-Based Emotion Recognition Using Deep Learning Approach

  2022cited by this paper
• GoEmotions: A Dataset of Fine-Grained Emotions

  2020cited by this paper
• A Transformer-based joint-encoding for Emotion Recognition and Sentiment Analysis

  2020cited by this paper
• Convolutional Recurrent Deep Learning Model for Sentence Classification

  2018cited by this paper
• Seernet at EmoInt-2017: Tweet Emotion Intensity Estimator

  2017cited by this paper
• Attention-Based Convolutional Neural Networks for Sentence Classification

  2016cited by this paper
• Recurrent Convolutional Neural Networks for Text Classification

  2015cited by this paper
• A Sensitivity Analysis of (and Practitioners’ Guide to) Convolutional Neural Networks for Sentence Classification

  2015cited by this paper
• Convolutional Neural Networks for Sentence Classification

  2014cited by this paper
• Learning Phrase Representations using RNN Encoder–Decoder for Statistical Machine Translation

  2014cited by this paper

• No citing papers are available for this paper.