講演抄録/キーワード |
講演名 |
2021-07-24 10:20
[ショートペーパー]COVID-19 and biased information dissemination on Twitter ○Gefei Li(AIST/Waseda Univ.)・Yijun Duan・Taehoon Kim・Kyoungsook Kim(AIST) DE2021-4 |
抄録 |
(和) |
(まだ登録されていません) |
(英) |
This study aims to examine if Twitter users with a higher level of similarity would have a more similar distribution of biased tweets. We randomly selected 100 users from an open-access COVID-19 Tweets dataset and fetched their Twitter timelines. We pre-trained the attention-based bidirectional long short-term memory (BiLSTM) model with a Media Bias Annotation Dataset and fine-tuned this model with 3,000 manually labelled Tweets concerning coronavirus-related topics. This model was used to classify users’ Tweets and returned the distribution of biased Tweets on their timelines. Users’ similarity was measured from two aspects: profile and text similarity. Ordinary Least Squares regression (OLS) correlation analysis suggests that both profile and text similarity are statistically significant in estimating the similarity of the distribution of biased tweets on users’ timelines. |
キーワード |
(和) |
/ / / / / / / |
(英) |
Bias / Twitter / Social Media / Profile Similarity / Text Similarity / NLP / / |
文献情報 |
信学技報, vol. 121, no. 125, DE2021-4, pp. 18-21, 2021年7月. |
資料番号 |
DE2021-4 |
発行日 |
2021-07-17 (DE) |
ISSN |
Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
DE2021-4 |