Weibo Rumor Detection Method Based
on User and Content Relationship
Zhongyue Zhou, Haijun Zhang
(&)
, and Weimin Pan
School of Computer Science and Technology, Xinjiang Normal University,
Urumqi 830054, China
zhjlp@163.com
Abstract. In order to effectively identify the rumor information in the Weibo
platform, we propose a combined model based on deep learning, which includes
convolutional neural network (CNN) that incorporates the attention mechanism
and combines with the neural network of long short-term memory (LSTM) to
implement a microblog rumor detection method for the characteristics of usercontent relations. Firstly, the convolutional neural network incorporating the
attention mechanism is used to extract the fine-grained features of the usercontent relationship. Secondly, the LSTM network is used for coarse-grained
feature extraction. Finally, the extracted feature vectors are classified by the
Softmax classifier so as to achieve a good effect of rumor detection.
Keywords: Weibo Á LSTM Á CNN Á Attention mechanism Á Softmax
1 Introduction
With the further development of the Internet era, social media has gradually increased
its position in people’s life, study, and work. Weibo is a broadcast social network
platform that spreads short-text real-time information through interactive methods. As
an open platform, Weibo has the characteristics of real-time, convenience, and rapid
dissemination in information. However, Weibo is not only an effective platform for
people to obtain information, but also an important tool for the breeding and dissemination of rumors.
At present, most of the rumors research based on Weibo are still using traditional
machine learning methods. Most of these methods use texts, user behaviors, time
series, and emotional values as rumor judgement basis. This paper leads into a
framework of deep learning, using a convolutional neural network model with attention
mechanism and a time-characteristic LSTM network model to effectively identify the
microblog rumors by using the relationship between user features and content features.
The main structure of this paper includes the following parts: The second part
introduces the related work on rumor recognition; the third part describes the rumor
detection method based on composite model of user-content relations; the fourth part
summarizes the proposed method.
© Springer Nature Singapore Pte Ltd. 2020
Q. Liang et al. (Eds.): Artificial Intelligence in China, LNEE 572, pp. 431–434, 2020.
https://doi.org/10.1007/978-981-15-0187-6_51
on User and Content Relationship
Zhongyue Zhou, Haijun Zhang
(&)
, and Weimin Pan
School of Computer Science and Technology, Xinjiang Normal University,
Urumqi 830054, China
zhjlp@163.com
Abstract. In order to effectively identify the rumor information in the Weibo
platform, we propose a combined model based on deep learning, which includes
convolutional neural network (CNN) that incorporates the attention mechanism
and combines with the neural network of long short-term memory (LSTM) to
implement a microblog rumor detection method for the characteristics of usercontent relations. Firstly, the convolutional neural network incorporating the
attention mechanism is used to extract the fine-grained features of the usercontent relationship. Secondly, the LSTM network is used for coarse-grained
feature extraction. Finally, the extracted feature vectors are classified by the
Softmax classifier so as to achieve a good effect of rumor detection.
Keywords: Weibo Á LSTM Á CNN Á Attention mechanism Á Softmax
1 Introduction
With the further development of the Internet era, social media has gradually increased
its position in people’s life, study, and work. Weibo is a broadcast social network
platform that spreads short-text real-time information through interactive methods. As
an open platform, Weibo has the characteristics of real-time, convenience, and rapid
dissemination in information. However, Weibo is not only an effective platform for
people to obtain information, but also an important tool for the breeding and dissemination of rumors.
At present, most of the rumors research based on Weibo are still using traditional
machine learning methods. Most of these methods use texts, user behaviors, time
series, and emotional values as rumor judgement basis. This paper leads into a
framework of deep learning, using a convolutional neural network model with attention
mechanism and a time-characteristic LSTM network model to effectively identify the
microblog rumors by using the relationship between user features and content features.
The main structure of this paper includes the following parts: The second part
introduces the related work on rumor recognition; the third part describes the rumor
detection method based on composite model of user-content relations; the fourth part
summarizes the proposed method.
© Springer Nature Singapore Pte Ltd. 2020
Q. Liang et al. (Eds.): Artificial Intelligence in China, LNEE 572, pp. 431–434, 2020.
https://doi.org/10.1007/978-981-15-0187-6_51