DOI QR코드

DOI QR Code

RNN과 트랜스포머 기반 모델들의 한국어 리뷰 감성분류 비교

Comparison of Sentiment Classification Performance of for RNN and Transformer-Based Models on Korean Reviews

  • 이재홍 (전남도립대학교 보건의료과)
  • Jae-Hong Lee (Dept. of Health & Medical Science, Jeonnam State University)
  • 투고 : 2023.06.27
  • 심사 : 2023.08.17
  • 발행 : 2023.08.31

초록

텍스트 문서에서 주관적인 의견과 감정을 긍정 혹은 부정으로 분류하고 식별하는 자연어 처리의 한 분야인 감성 분석은 고객 선호도 분석을 통해 다양한 홍보 및 서비스에 활용할 수 있다. 이를 위해 최근 머신러닝과 딥러닝의 다양한 기법을 활용한 연구가 진행되어 왔다. 본 연구에서는 기존의 RNN 기반 모델들과 최근 트랜스포머 기반 언어 모델들을 활용하여 영화, 상품 및 게임 리뷰를 대상으로 감성 분석의 정확도를 비교 분석하여 최적의 언어 모델을 제안하고자 한다. 실험 결과 한국어 말뭉치로 사전 학습된 모델들 중 LMKor-BERT와 GPT-3가 상대적으로 좋은 정확도를 보여주었다.

Sentiment analysis, a branch of natural language processing that classifies and identifies subjective opinions and emotions in text documents as positive or negative, can be used for various promotions and services through customer preference analysis. To this end, recent research has been conducted utilizing various techniques in machine learning and deep learning. In this study, we propose an optimal language model by comparing the accuracy of sentiment analysis for movie, product, and game reviews using existing RNN-based models and recent Transformer-based language models. In our experiments, LMKorBERT and GPT3 showed relatively good accuracy among the models pre-trained on the Korean corpus.

키워드

참고문헌

  1. J.-H., Seo, "Image Classification, Deep Learning, Convolutional Neural Network, Transfer Learning," J. of The Korea Institute of Electronic Communication Sciences, vol. 18, no. 3, 2023, pp. 413-420.
  2. J.-M. Jo, "Time Series Data Processing Deep Learning system for Prediction of Hospital Outpatient Number," J. of The Korea Institute of Electronic Communication Sciences, vol. 16 no. 2, 2021, pp. 313-318.
  3. Y.-T. Oh, M.-T. Kim, and W.-J. Kim, "Korean Movie-review Sentiment Analysis Using Parallel Stacked Bidirectional LSTM," J. of The Korean Institute of Information Scientists and Engineers, vol. 46, no. 1, 2019, pp. 45-49. https://doi.org/10.5626/JOK.2019.46.1.45
  4. C.-G. Lim, I. Aliyu, and R.-M. Mahmood, "LSTM Hyperparameter Optimization for an EEG-Based Efficient Emotion Classification in BCI," J. of The Korea Institute of Electronic Communication Sciences, vol. 14, no. 6, 2019, pp. 1171-1180.
  5. J.-W. Min, S.-H. Na, J.-H. Shin, and Y.-K. Kim, "RoBERTa for Korean Natural Language Processing: Named Entity Recognition, Sentiment Analysis, Dependency Parsing," Proc. of Korea Software Congress 2019, , 2019, pp. 407-409.
  6. T.-H. Kim, D.-B. Cho, H.-Y. Lee, H.-J. Won, and S.-S. Kang, "Sentiment Analysis System by Using BERT Language Model," Proc. of Annual Conference, Online, Korea, 2020, pp. 975-977.
  7. M.-H. Lee, "Text sentiment analysis using deep learning and ensemble technique," Proc. of Korea Computer Congress 2021, Jeju, Korea, 2021, pp. 451-453.
  8. S.-H. Hwang, "Text sentiment Analysis Based on Transformer Models using an emotional dictionary," Proc. of Korea Computer Congress 2021, Jeju, Korea, 2021, pp. 876-878.
  9. K.-R. Lim and H.-S. Lim, "Comparative Analysis of Emotional Classification Performance by Domain between Korean Natural Language Processing Models," Proc. of Korea Computer Congress 2022, Jeju, Korea, 2022, pp. 332-334.
  10. S. Ravichandiran, Getting Started with Google BERT: Build and train state-of-the-art natural language processing models using BERT. Birmingham: Packt Publishing, 2021.