Rumor Detection over Varying Time Windows

Kwon, Sejeong; Cha, Meeyoung; Jung, Kyomin

doi:10.1371/journal.pone.0168344

Cited by 286 publications

(162 citation statements)

References 44 publications

(27 reference statements)

Supporting

Mentioning

157

Contrasting

Order By: Relevance

“…However, the most important (and interesting) cases are when the delay in that decision could also have negative or risky implications. This scenario, known as "early risk detection" have gained increasing interest in recent years with potential applications in rumor detection (Ma et al, 2015(Ma et al, , 2016Kwon et al, 2017), sexual predator detection and aggressive text identification (Escalante et al, 2017), depression detection (Losada et al, 2017;Losada & Crestani, 2016) or terrorism detection (Iskandar, 2017).…”

Section: Analysis Of Sequential Data: Early Classificationmentioning

confidence: 99%

A text classification framework for simple and effective early depression detection over social media streams

Burdisso

Errecalde

Montes-y-Gómez

2019

Expert Systems with Applications

124

View full text Add to dashboard Cite

With the rise of the Internet, there is a growing need to build intelligent systems that are capable of efficiently dealing with early risk detection (ERD) problems on social media, such as early depression detection, early rumor detection or identification of sexual predators. These systems, nowadays mostly based on machine learning techniques, must be able to deal with data streams since users provide their data over time. In addition, these systems must be able to decide when the processed data is sufficient to actually classify users. Moreover, since ERD tasks involve risky decisions by which people's lives could be affected, such systems must also be able to justify their decisions. However, most standard and state-of-the-art supervised machine learning models (such as SVM, MNB, Neural Networks, etc.) are not well suited to deal with this scenario. This is due to the fact that they either act as black boxes or do not support incremental classification/learning. In this paper we introduce SS3, a novel supervised learning model for text classification that naturally supports these aspects. SS3 was designed to be used as a general framework to deal with ERD problems. We evaluated our model on the CLEF's eRisk2017 pilot task on early depression detection. Most of the 30 contributions submitted to this competition used state-of-the-art methods. Experimental results show that our classifier was able to outperform these models and standard classifiers, despite being less computationally expensive and having the ability to explain its rationale.

show abstract

Section: Analysis Of Sequential Data: Early Classificationmentioning

confidence: 99%

A text classification framework for simple and effective early depression detection over social media streams

Burdisso

Errecalde

Montes-y-Gómez

2019

Expert Systems with Applications

124

View full text Add to dashboard Cite

show abstract

“…Tracing the diffusion path and inferring the network structure are two expensive features that cannot be obtained in the early stage of news propagation, while using textual and user sequence features act consistently good even in the beginning of diffusion, when the information is partially available [28].…”

Section: B Model Descriptionmentioning

confidence: 99%

News labeling as early as possible

Ramezani

Rafiei

Soroush

et al. 2019

Proceedings of the 2019 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining

View full text Add to dashboard Cite

Making disguise between real and fake news propagation through online social networks is an important issue in many applications. The time gap between the news release time and detection of its label is a significant step towards broadcasting the real information and avoiding the fake. Therefore, one of the challenging tasks in this area is to identify fake and real news in early stages of propagation. However, there is a tradeoff between minimizing the time gap and maximizing accuracy. Despite recent efforts in detection of fake news, there has been no significant work that explicitly incorporates early detection in its model. In this paper, we focus on accurate early labeling of news, and propose a model by considering earliness both in modeling and prediction. The proposed method utilizes recurrent neural networks with a novel loss function, and a new stopping rule. Given the context of news, we first embed it with a class-specific text representation. Then, we utilize the available public profile of users, and speed of news diffusion, for early labeling of the news. Experiments on real datasets demonstrate the effectiveness of our model both in terms of early labelling and accuracy, compared to the state of the art baseline and models.

show abstract

“…Spreading dynamics is an important issue in spreading and controlling [1][2][3] of 2 rumor [4][5][6][7] and disease [8][9][10][11], marketing [12], recommending [13][14][15], source 3 detecting [16,17], and many other interesting topics [18][19][20][21][22]. How to predict the 4 infection probability [23], infected scale [24,25], and even the infected nodes precisely 5 has been gotten much attention in recent years.…”

Section: Introductionmentioning

confidence: 99%

Spreading predictability in complex networks

Chen

Zhao

Wang

et al. 2020

Preprint

View full text Add to dashboard Cite

Spreading dynamics analysis is an important and interesting topic since it has many applications such as rumor or disease controlling, viral marketing and information recommending. Many state-of-the-art researches focus on predicting infection scale or threshold. Few researchers pay attention to the predicting of infection nodes from a snapshot. With developing of precision marketing, recommending and, controlling, how to predict infection nodes precisely from snapshot becomes a key issue in spreading dynamics analysis. In this paper, a probability based prediction model is presented so as to estimate the infection nodes from a snapshot of spreading. Experimental results on synthetic and real networks demonstrate that the model proposed could predict the infection nodes precisely in the sense of probability.Researchers have gotten many achievements on macro level of spreading such as 7 phase transition of spreading [26] and basic reproduction number [27]. Up to now, 8 many researches focus on estimating of infection scale. The simplest one is mean-field 9 model, in which, the spreading coverage can be predicted by using differential 10 equations [24]. Besides mean-field model, some more realistic models such as pair 11 approximation [25] and permutation entropy [28] are considered to predict the 12infection scale or infectious disease outbreaks. The main difference between mean-field 13 and pair approximation is that the former(latter) approximates high-order moments in 14 term of first (second) order ones. In [28], the researchers studied the predictability of a 15

show abstract

Rumor Detection over Varying Time Windows

Cited by 286 publications

References 44 publications

A text classification framework for simple and effective early depression detection over social media streams

A text classification framework for simple and effective early depression detection over social media streams

News labeling as early as possible

Spreading predictability in complex networks

Contact Info

Product

Resources

About