Journal of the Korea Institute of Information and Communication Engineering (한국정보통신학회논문지)

The Korea Institute of Information and Commucation Engineering (한국정보통신학회)
권호 표지
DOI QR코드

DOI QR Code

Analyzing Media Bias in News Articles Using RNN and CNN

순환 신경망과 합성곱 신경망을 이용한 뉴스 기사 편향도 분석

  • Received : 2020.05.13
  • Accepted : 2020.06.03
  • Published : 2020.08.31
Download PDF
⟨ Previous Next ⟩

Abstract

While search portals' 'Portal News' account for the largest portion of aggregated news outlet, its neutrality as an outlet is questionable. This is because news aggregation may lead to prejudiced information consumption by recommending biased news articles. In this paper we introduce a new method of measuring political bias of news articles by using deep learning. It can provide its readers with insights on critical thinking. For this method, we build the dataset for deep learning by analyzing articles' bias from keywords, sourced from the National Assembly proceedings, and assigning bias to said keywords. Based on these data, news article bias is calculated by applying deep learning with a combination of Convolution Neural Network and Recurrent Neural Network. Using this method, 95.6% of sentences are correctly distinguished as either conservative or progressive-biased; on the entire article, the accuracy is 46.0%. This enables analyzing any articles' bias between conservative and progressive unlike previous methods that were limited on article subjects.

오늘날의 검색 포털은 뉴스의 창구로서는 가장 큰 비율을 차지하지만, 중립성에 대해서는 의문이 제기되고 있다. 이는 포털 뉴스가 편향된 정보의 소비를 유도할 수 있기 때문이다. 본 논문은 뉴스 기사의 정치적 편향도를 딥러닝을 이용하여 측정하는 방법에 대하여 소개한다. 이는 기사를 비판적으로 바라보는 시각을 뉴스 독자에게 제공할 것이다. 구체적으로, 국회 회의록에서 추출한 키워드에 편향도를 부여하고, 이를 기반으로 기사의 편향도를 분석하여 머신러닝용 데이터를 구축하였다. 최종적으로 순환 신경망과 합성곱 신경망을 융합한 딥러닝을 통해 기사의 편향도를 계산하는 것을 목표로 하였다. 학습한 모델의 정확도를 분석한 결과 문장별 편향의 좌/우편향 판정은 95.6%의 정확도를 보였으나, 신문기사 전체에서는 46.0%의 정확도를 보였다. 이는 기존의 여러 편향성 연구와 다르게 특정 주제에 한정되지 않고 기사의 보수-진보 편향성을 분석할 수 있도록 한다.

Keywords

References

  1. S. J. Kim, and Y. G. Jeong, "A Study on the Characteristics of Political Bias of Korean Press : Focused on the Analysis of 19th Presidential Election Coverage," Korean Journal of Communication & Information, vol. 88, pp.110-145, 2018. https://doi.org/10.46407/kjci.2018.04.88.110
  2. T. Hamborg, K. Donnay, and B. Gipp, Automated identification of media bias in news articles: an interdisciplinary literature review. International Journal on Digital Libraries, vol. 20, no. 4, 391-415. 2019. https://doi.org/10.1007/s00799-018-0261-y
  3. A. Balahur, R. Steinberger, M. Kabadjov, V. Zavarella, E. Goot, M. Halkia, B. Pouliquen, and J. Belyaeva, Sentiment analysis in the news. arXiv Prepr. arXiv1309.6202 (2013).
  4. Chosun News. [20th National Assembly Ideology Map] Comparing the 17-20th National Assembly Ideology [Internet]. Available: https://news.chosun.com/site/data/html_dir/2018/01/08/20180108010 43.html.
  5. M. Gentzkow, and J. M. Shapiro, "What drives media slant? Evidence from U.S. daily newspapers," Econometrica, vol. 78, no.1, pp. 35-71, 2010. https://doi.org/10.3982/ECTA7195
  6. D. W. Choi, "Internet Portal Competition and Economic Incentives to Tailor News Slant," The Korean Journal of Industrial Organization, vol. 25, no. 2, pp. 1-40, Jun. 2017. https://doi.org/10.36354/KJIO.25.2.1
  7. Tensorflow. tf.keras.preprocessing.text.Tokenizer[Internet]. Available:https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/text/Tokenizer.
  8. S. Y. Hong, S. H. Na, J. H. Shin, and Y. K. Kim, "BERT and ELMo for contextualized word embeddings in Korean Dependency Parsing," The Korean Institute of Information Scientists and Engineers, 2019.6, 491-493(3 pages).