Rumor Detection on Hierarchical Attention Network with User and Sentiment Information
Social media has developed rapidly due to its openness and freedom, and people can post information on Internet anytime and anywhere. However, social media has also become the main way for rumors to spread largely and quickly. Hence, it has become a huge
- PDF / 640,601 Bytes
- 12 Pages / 439.37 x 666.142 pts Page_size
- 40 Downloads / 221 Views
[email protected], {qianzhong,pfli,xiaoxzhu,qmzhu}@suda.edu.cn 2 AI Research Institute, Soochow University, Suzhou, China
Abstract. Social media has developed rapidly due to its openness and freedom, and people can post information on Internet anytime and anywhere. However, social media has also become the main way for rumors to spread largely and quickly. Hence, it has become a huge challenge to automatically detect rumors among such a huge amount of information. Currently, there are many neural network methods, which mainly considered text features but did not pay enough attention to user and sentiment information that are also useful clues for rumor detection. Therefore, this paper proposes a hierarchical attention network with user and sentiment information (HiAN-US) for rumor detection, which first uses the transformer encoder to learn the semantic information at both word-level and tweet-level, then integrates user and sentiment information via attention mechanism. Experiments on the Twitter15, Twitter16 and PHEME datasets show that our model is more effective than several state-of-the-art baselines. Keywords: Rumor detection · User and sentiment information · Hierarchical attention network
1 Introduction Rumors, usually used to spread panic and confusion, are untrue or inaccurate information breed on public platforms and Rumor Detection (RD) is to judge whether the information is true or false. Our work focuses on using relative public information to detect the false information spreading on social media. The key behind this work is that users on social media can express their opinions on the information disseminated on social media, and can provide evidence and speculation on false information [1]. In recent years, it has become increasingly popular to use neural network models to detect rumors. By modeling text information on social media, for example, Ma et al. [6–10] proposed a series of RNN-based methods, and these methods can automatically obtain a high-level text representation to detect the true degree of information. However, they only focused on how to use the text information of the rumor, and did not pay enough attention to user information, or even ignored it. Moreover, these methods hardly considered the role of sentiment information. Different users hold different degrees of credibility, and the sentiment expressed by them is directly related to their opinions. © Springer Nature Switzerland AG 2020 X. Zhu et al. (Eds.): NLPCC 2020, LNAI 12431, pp. 366–377, 2020. https://doi.org/10.1007/978-3-030-60457-8_30
Rumor Detection on Hierarchical Attention Network
367
Table 1. An example of a rumor source tweet and its user features Source tweet
Breaking! A four meter long Cobra with three heads was found in Myanmar!
User Features
username: ABCD_1234 verified: False description: follower: 15 listed_count: 300 user_creat_time: 2011/10/4 9:36:17 tweet_creat_time: 2011/10/4 17:52:36 ……
Table 1 shows an example of a rumor source tweet and its user features. User ABCD_1234 posted a tweet: A 4 m lo
Data Loading...
