Solving the Problem of Classifying the Emotional Tone of a Message with Determining the Most Appropriate Neural Network Architecture
https://doi.org/10.26907/1562-5419-2023-26-4-396-413
Abstract
To determine the most effective approach for solving the task of classifying the emotional tone of a message, we trained selected neural network models on various sets of training data. Next, based on the performance metric of the percentage of correctly classified responses on a test data set, we compared combinations of training data sets and various models trained on them. During the writing of this article, we trained four neural network models on three different sets of training data. By comparing the accuracy of the responses from each model trained on different training data sets, conclusions were drawn regarding the neural network model best suited for solving the task at hand.
About the Authors
D. I. Bagautdinov
Kazan (Volga region) Federal University
Russian Federation
Russian Federation
S. Salman
Kazan (Volga region) Federal University
Russian Federation
Russian Federation
V. A. Alekseev
Kazan (Volga region) Federal University
Russian Federation
Russian Federation
R. M. Usmonov
Kazan (Volga region) Federal University
Russian Federation
Russian Federation
References
1. Gosai D. D., Gohil H. J., Jayswal H. S. A Review on a Emotion Detection and Recognization from Text Using Natural Language Processing. URL: http://www.ripublication.com/ijaer18/ijaerv13n9_33.pdf
2. Acheampong F. A., Nunoo-Mensah H., Wenyu Chen. Transformer models for text-based emotion detection: a review of BERT-based approaches. URL: https://link.springer.com/article/10.1007/s10462-021-09958-2
3. Naresh S. S., Khethawat S. Emotion Detection from Text URL: https://arxiv.org/abs/1205.4944
4. Baccianella S., Esuli A., Sebastiani F. Sentiwordnet 3.0: an enhanced lexical resource for sentiment analysis and opinion mining // Lrec. 2010. P. 2200–2204
5. Akbik A., Blythe D., Vollgraf R. Contextual string embeddings for sequence labeling // Proceedings of the 27th international conference on computational linguistics. 2010. P. 1638–1649.
6. Keras. URL: https://keras.io/
7. Tensorflow. URL: https://www.tensorflow.org/
8. Numpy. URL: https://numpy.org/
9. Pandas. URL: https://pandas.pydata.org/
10. Aiogram. URL: https://aiogram.dev
11. Asyncio. URL: https://docs.python.org/3/library/asyncio.html
12. Aiohttp. URL: https://docs.aiohttp.org/en/stable/index.html
13. GloVe: Global Vectors for Word Representation. URL: https://nlp.stanford.edu/projects/glove/
14. Google Collab. URL: https://colab.research.google.com
15. Jupyter Notebook. URL: https://docs.jupyter.org/en/latest/install/notebook-classic.html
16. Multi-Class Text Emotion Analysis. URL: https://github.com/kanchitank/Text-Emotion-Analysis
17. Sentiment140 dataset with 1.6 million tweets. URL: https://www.kaggle.com/datasets/kazanova/sentiment140
18. Twitter Emotion Classification. URL: https://www.kaggle.com/code/shtrausslearning/twitter-emotion-classification?scriptVersionId=126983369
19. BiLSTM. URL: https://www.kaggle.com/code/mlwhiz/bilstm-pytorch-and-keras
Review
For citations:
Bagautdinov D.I., Salman S., Alekseev V.A., Usmonov R.M. Solving the Problem of Classifying the Emotional Tone of a Message with Determining the Most Appropriate Neural Network Architecture. Russian Digital Libraries Journal. 2023;26(4):396-413. (In Russ.) https://doi.org/10.26907/1562-5419-2023-26-4-396-413
Views:
26
Email this article 