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

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.

[1]  Mann Oo. Hay Emotion recognition in human-computer interaction , 2012 .

[2]  Xiao-Ying Liu,et al.  Measuring Semantic Similarity in Wordnet , 2007, 2007 International Conference on Machine Learning and Cybernetics.

[3]  Prema Nedungadi,et al.  Hybrid Approach for Emotion Classification of Audio Conversation Based on Text and Speech Mining , 2015 .

[4]  Chung-Hsien Wu,et al.  Detecting Unipolar and Bipolar Depressive Disorders from Elicited Speech Responses Using Latent Affective Structure Model , 2020, IEEE Transactions on Affective Computing.

[5]  Geoffrey E. Hinton,et al.  Autoencoders, Minimum Description Length and Helmholtz Free Energy , 1993, NIPS.

[6]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[7]  Lung-Hao Lee,et al.  Building Chinese Affective Resources in Valence-Arousal Dimensions , 2016, NAACL.

[8]  Muhammad Anwarul Azim,et al.  Introducing active learning on Text to Emotion Analyzer , 2014, 2014 17th International Conference on Computer and Information Technology (ICCIT).

[9]  Astrid Paeschke,et al.  A database of German emotional speech , 2005, INTERSPEECH.

[10]  Chung-Hsien Wu,et al.  Emotion Recognition of Affective Speech Based on Multiple Classifiers Using Acoustic-Prosodic Information and Semantic Labels , 2015, IEEE Transactions on Affective Computing.

[11]  Andrés Montoyo,et al.  Detecting implicit expressions of emotion in text: A comparative analysis , 2012, Decis. Support Syst..

[12]  Chung-Hsien Wu,et al.  Coupled HMM-based multimodal fusion for mood disorder detection through elicited audio–visual signals , 2016, J. Ambient Intell. Humaniz. Comput..

[13]  Chung-Hsien Wu,et al.  Emotion recognition from text using semantic labels and separable mixture models , 2006, TALIP.

[14]  Lexical Semantic Representation and Semantic Composition , 2016 .

[15]  Von-Wun Soo,et al.  Towards Text-based Emotion Detection A Survey and Possible Improvements , 2009, 2009 International Conference on Information Management and Engineering.