LSTM-based Text Emotion Recognition Using Semantic and Emotional Word Vectors

Ming Hsiang Su, Chung-Hsien Wu, Kun Yi Huang, Qian Bei Hong

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

This study proposes a long-short term memory (LSTM)-based approach to text emotion recognition based on semantic word vector and emotional word vector of the input text. For each word in an input text, the semantic word vector is extracted from the word 2vec model. Besides, each lexical word is projected to all the emotional words defined in an affective lexicon to derive an emotional word vector. An autoencoder is then adopted to obtain the bottleneck features from the emotional word vector for dimensionality reduction. The autoencoder bottleneck features are then concatenated with the features in the semantic word vector to form the final textual features for emotion recognition. Finally, given the textual feature sequence of the entire sentence, the LSTM is used for emotion recognition by modeling the contextual emotion evolution of the input text. For evaluation, the NLPCC-MHMC-TE database containing seven emotion categories: anger, boredom, disgust, anxiety, happiness, sadness, and surprise was constructed and used. Five-fold cross-validation was employed to evaluate the performance of the proposed method. Experimental results show that the proposed LSTM-based method achieved a recognition accuracy of 70.66%, improving 5.33% compared with the CNN-based method. Besides, the proposed method based on integration of the semantic word vector and emotional word vector of the input text outperformed that using the individual feature vector.

Original languageEnglish
Title of host publication2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781538653111
DOIs
Publication statusPublished - 2018 Sep 21
Event1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018 - Beijing, China
Duration: 2018 May 202018 May 22

Publication series

Name2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018

Other

Other1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018
CountryChina
CityBeijing
Period18-05-2018-05-22

Fingerprint

Semantics
Long short-term memory

All Science Journal Classification (ASJC) codes

  • Computer Vision and Pattern Recognition
  • Human-Computer Interaction
  • Signal Processing
  • Artificial Intelligence

Cite this

Su, M. H., Wu, C-H., Huang, K. Y., & Hong, Q. B. (2018). LSTM-based Text Emotion Recognition Using Semantic and Emotional Word Vectors. In 2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018 [8470378] (2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ACIIAsia.2018.8470378
Su, Ming Hsiang ; Wu, Chung-Hsien ; Huang, Kun Yi ; Hong, Qian Bei. / LSTM-based Text Emotion Recognition Using Semantic and Emotional Word Vectors. 2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018. Institute of Electrical and Electronics Engineers Inc., 2018. (2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018).
@inproceedings{6535648005b3464bb1c0de3f6d23b9f7,
title = "LSTM-based Text Emotion Recognition Using Semantic and Emotional Word Vectors",
abstract = "This study proposes a long-short term memory (LSTM)-based approach to text emotion recognition based on semantic word vector and emotional word vector of the input text. For each word in an input text, the semantic word vector is extracted from the word 2vec model. Besides, each lexical word is projected to all the emotional words defined in an affective lexicon to derive an emotional word vector. An autoencoder is then adopted to obtain the bottleneck features from the emotional word vector for dimensionality reduction. The autoencoder bottleneck features are then concatenated with the features in the semantic word vector to form the final textual features for emotion recognition. Finally, given the textual feature sequence of the entire sentence, the LSTM is used for emotion recognition by modeling the contextual emotion evolution of the input text. For evaluation, the NLPCC-MHMC-TE database containing seven emotion categories: anger, boredom, disgust, anxiety, happiness, sadness, and surprise was constructed and used. Five-fold cross-validation was employed to evaluate the performance of the proposed method. Experimental results show that the proposed LSTM-based method achieved a recognition accuracy of 70.66{\%}, improving 5.33{\%} compared with the CNN-based method. Besides, the proposed method based on integration of the semantic word vector and emotional word vector of the input text outperformed that using the individual feature vector.",
author = "Su, {Ming Hsiang} and Chung-Hsien Wu and Huang, {Kun Yi} and Hong, {Qian Bei}",
year = "2018",
month = "9",
day = "21",
doi = "10.1109/ACIIAsia.2018.8470378",
language = "English",
series = "2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
booktitle = "2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018",
address = "United States",

}

Su, MH, Wu, C-H, Huang, KY & Hong, QB 2018, LSTM-based Text Emotion Recognition Using Semantic and Emotional Word Vectors. in 2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018., 8470378, 2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018, Institute of Electrical and Electronics Engineers Inc., 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018, Beijing, China, 18-05-20. https://doi.org/10.1109/ACIIAsia.2018.8470378

LSTM-based Text Emotion Recognition Using Semantic and Emotional Word Vectors. / Su, Ming Hsiang; Wu, Chung-Hsien; Huang, Kun Yi; Hong, Qian Bei.

2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018. Institute of Electrical and Electronics Engineers Inc., 2018. 8470378 (2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - LSTM-based Text Emotion Recognition Using Semantic and Emotional Word Vectors

AU - Su, Ming Hsiang

AU - Wu, Chung-Hsien

AU - Huang, Kun Yi

AU - Hong, Qian Bei

PY - 2018/9/21

Y1 - 2018/9/21

N2 - This study proposes a long-short term memory (LSTM)-based approach to text emotion recognition based on semantic word vector and emotional word vector of the input text. For each word in an input text, the semantic word vector is extracted from the word 2vec model. Besides, each lexical word is projected to all the emotional words defined in an affective lexicon to derive an emotional word vector. An autoencoder is then adopted to obtain the bottleneck features from the emotional word vector for dimensionality reduction. The autoencoder bottleneck features are then concatenated with the features in the semantic word vector to form the final textual features for emotion recognition. Finally, given the textual feature sequence of the entire sentence, the LSTM is used for emotion recognition by modeling the contextual emotion evolution of the input text. For evaluation, the NLPCC-MHMC-TE database containing seven emotion categories: anger, boredom, disgust, anxiety, happiness, sadness, and surprise was constructed and used. Five-fold cross-validation was employed to evaluate the performance of the proposed method. Experimental results show that the proposed LSTM-based method achieved a recognition accuracy of 70.66%, improving 5.33% compared with the CNN-based method. Besides, the proposed method based on integration of the semantic word vector and emotional word vector of the input text outperformed that using the individual feature vector.

AB - This study proposes a long-short term memory (LSTM)-based approach to text emotion recognition based on semantic word vector and emotional word vector of the input text. For each word in an input text, the semantic word vector is extracted from the word 2vec model. Besides, each lexical word is projected to all the emotional words defined in an affective lexicon to derive an emotional word vector. An autoencoder is then adopted to obtain the bottleneck features from the emotional word vector for dimensionality reduction. The autoencoder bottleneck features are then concatenated with the features in the semantic word vector to form the final textual features for emotion recognition. Finally, given the textual feature sequence of the entire sentence, the LSTM is used for emotion recognition by modeling the contextual emotion evolution of the input text. For evaluation, the NLPCC-MHMC-TE database containing seven emotion categories: anger, boredom, disgust, anxiety, happiness, sadness, and surprise was constructed and used. Five-fold cross-validation was employed to evaluate the performance of the proposed method. Experimental results show that the proposed LSTM-based method achieved a recognition accuracy of 70.66%, improving 5.33% compared with the CNN-based method. Besides, the proposed method based on integration of the semantic word vector and emotional word vector of the input text outperformed that using the individual feature vector.

UR - http://www.scopus.com/inward/record.url?scp=85055490722&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85055490722&partnerID=8YFLogxK

U2 - 10.1109/ACIIAsia.2018.8470378

DO - 10.1109/ACIIAsia.2018.8470378

M3 - Conference contribution

T3 - 2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018

BT - 2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Su MH, Wu C-H, Huang KY, Hong QB. LSTM-based Text Emotion Recognition Using Semantic and Emotional Word Vectors. In 2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018. Institute of Electrical and Electronics Engineers Inc. 2018. 8470378. (2018 1st Asian Conference on Affective Computing and Intelligent Interaction, ACII Asia 2018). https://doi.org/10.1109/ACIIAsia.2018.8470378