ABSTRACT
Ensemble learning can train and combine multiple classifiers where the predictions are used as new features to train a meta-classifier. This improves the accuracy of the model. This paper proposes a multi granularity model based on Stacking ensemble learning for Korean text classification. Firstly, eojeol and subeojeol granularity is proposed according to the Korean language composition. Since different feature granularity contains different semantic information, compare the six different granularities of the phoneme, syllable, subword, word, subeojeol, and eojeol in Korean text classification task. Secondly, construct suffix words based on Korean grammatical morphology and compare the different granularities effects after suffix preprocessing. Finally, propose a multi granularity ensemble learning model based on Korean called MGEL-K. To enrich the diversity of ensemble learning using different granularities, making differences between learners. The results show that MGEL-K model proposed in this paper works best in the Korean text classification task with an accuracy of 92.33%.
- L. K. Hansen and P. Salamon, “Neural network ensembles,” IEEE Trans. Pattern Anal. Mach. Intell., vol. 12, no. 10, pp. 993–1001, 1990. Google ScholarDigital Library
- T. G. Dietterich, “Ensemble methods in machine learning,” in International workshop on multiple classifier systems, 2000, pp. 1–15. Google ScholarDigital Library
- R. Caruana and A. Niculescu-Mizil, “An empirical comparison of supervised learning algorithms,” in Proceedings of the 23rd international conference on Machine learning, 2006, pp. 161–168 Google ScholarDigital Library
- X. Zhang, J. Zhao, and Y. LeCun, “Character-level convolutional networks for text classification,” in Advances in neural information processing systems, 2015, pp. 649–657. Google ScholarDigital Library
- R. Sennrich, B. Haddow, and A. Birch, “Neural machine translation of rare words with subword units,” arXiv Prepr. arXiv1508.07909, 2015.Google Scholar
- Mintae Kim, Yeongtaek Oh, and Wooju Kim, “Sentence similarity prediction based on siamese CNN-Bidirectional LSTM with Self-attention,” Korean Inst. Inf. Sci. Eng., vol. 46, no. 3, pp. 241–245, 2019.Google Scholar
- X. Chen, L. Xu, Z. Liu, M. Sun, and H. Luan, “Joint learning of character and word embeddings,” 2015.Google Scholar
- J. Yu, X. Jian, H. Xin, and Y. Song, “Joint embeddings of chinese words, characters, and fine-grained subcharacter components,” in Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, 2017, pp. 286–291.Google ScholarCross Ref
- X. Meng, Y. Zhao, and M. Fang, “Multilingual text classification method based on bi-directional long term memory and convolutional neural network,” Appl. Res. Comput., vol. 37, no. 9, pp. 2669–2673, 2020.Google Scholar
- E. L. Park and S. Cho, “KoNLPy: Korean natural language processing in Python,” Proc. 26th Annu. Conf. Hum. Cogn. Lang. Technol., pp. 133–136, 2014.Google Scholar
- T. Kudo, “Subword regularization: Improving neural network translation models with multiple subword candidates,” arXiv Prepr. arXiv1804.10959, 2018.Google Scholar
- J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “Bert: Pre-training of deep bidirectional transformers for language understanding,” arXiv Prepr. arXiv1810.04805, 2018.Google Scholar
- Cho S, Whitman J, “Korean: A Linguistic Introduction,” Cambridge University Press, 2019, pp. 31-35.Google Scholar
- F. Yang, Y. Zhao, R. Cui, and Z. Yi, “Words Alignment in Parallel Corpus Based on Translation Probability,” J. Chinese Inf. Process., vol. 33, no. 12, pp. 37–44, 2019.Google Scholar
- R. E. Schapire, “The strength of weak learnability,” Mach. Learn., vol. 5, no. 2, pp. 197–227, 1990. Google ScholarDigital Library
- L. Breiman, “Bagging predictors,” Mach. Learn., vol. 24, no. 2, pp. 123–140, 1996. Google ScholarDigital Library
- D. H. Wolpert, “Stacked generalization,” Neural networks, vol. 5, no. 2, pp. 241–259, 1992. Google ScholarDigital Library
- K. Tumer and J. Ghosh, “Analysis of decision boundaries in linearly combined neural classifiers,” Pattern Recognit., vol. 29, no. 2, pp. 341–348, 1996.Google ScholarCross Ref
- Y. Kim, “Convolutional neural networks for sentence classification,” arXiv Prepr. arXiv1408.5882, 2014.Google ScholarCross Ref
- A. Mnih and G. E. Hinton, “A scalable hierarchical distributed language model,” Adv. Neural Inf. Process. Syst., vol. 21, pp. 1081–1088, 2008. Google ScholarDigital Library
- Z. Lin , “A structured self-attentive sentence embedding,” arXiv Prepr. arXiv1703.03130, 2017.Google Scholar
- A. Vaswani , “Attention is all you need,” in Advances in neural information processing systems, 2017, pp. 5998–6008. Google ScholarDigital Library
- M. Tian, Y. Zhao, and R. Cui, “Identifying Word Translations in Scientific Literature Based on Labeled Bilingual Topic Model and Co-occurrence Features,” in Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data, Springer, 2018, pp. 76–87.Google ScholarCross Ref
Index Terms
- Research on Multi-granularity Ensemble Learning Based on Korean
Recommendations
TextCNN-based ensemble learning model for Japanese Text Multi-classification
AbstractIn this paper, we aim at improving Japanese text classification using TextCNN-based ensemble learning model. Specifically, we first construct three different sub-classifiers, combining ALBERT, RoBERTa, DistilBERT with TextCNN, respectively; and ...
Graphical abstractDisplay Omitted
Highlights- Three TextCNN-based sub-classifiers for Japanese text classification are designed.
- A Bagging ensemble learning model is proposed to combine three different subclassifiers for multi-label Japanese text classification.
- A Japanese ...
Ensemble of feature sets and classification algorithms for sentiment classification
In this paper, we make a comparative study of the effectiveness of ensemble technique for sentiment classification. The ensemble framework is applied to sentiment classification tasks, with the aim of efficiently integrating different feature sets and ...
Building Locally Discriminative Classifier Ensemble Through Classifier Fusion Among Nearest Neighbors
PCM 2016: 17th Pacific-Rim Conference on Advances in Multimedia Information Processing - Volume 9916Many studies on ensemble learning that combines multiple classifiers have shown that, it is an effective technique to improve accuracy and stability of a single classifier. In this paper, we propose a novel discriminative classifier fusion method, which ...
Comments