Towards Learning-Based, Content-Agnostic Detection of Social Bot Traffic

Feng, Yebo; Li, Jun; Jiao, Lei; Wu, Xintao

doi:10.1109/tdsc.2020.3047399

Cited by 9 publications

(5 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Due to the dataset imbalance problem, BinWu [5] et al proposed an improved conditional generative adversarial network to expand the dataset, that is, combining Wasserstein distance with gradient penalty and clustering algorithms as a new data augmentation method, In order to improve the accuracy of social bot detection, experiments show that the F1 score is 97.56%. Feng et al [6] proposed a method called BotFlowMon to detect social bot from the perspective of network data traffic. The experimental results show that BotFlowMon has an accuracy rate of 96.1% and can see social bot in an average of 0.71 seconds.…”

Section: Machine Learning Based Detectionmentioning

confidence: 99%

“…Can. (6) Remove stop words, and filter stop words with the help of the functions provided by the word segmentation library. Usually, stop words have less value to sentences, and filtering can improve the efficiency of the program.…”

Section: Work Processmentioning

confidence: 99%

“…Usually, tweets published by normal users are rarely retweeted by other users, so the ret value of social bot will be higher than that of normal users. (6) The original amount: The original proportion of user tweets. Normal users use Twitter to express their views on life dynamics or events, while social bot will retweet other people's tweets for this purpose, so usually The Originalp value for normal users will be higher.…”

Section: Figurementioning

confidence: 99%

See 2 more Smart Citations

Quantum-based Detection of Higly Semantically Similar Social bot

Liu

2023

FCIS

View full text Add to dashboard Cite

Social bot often publish content that has a high degree of similarity in text or semantics. Combining this feature, this paper designs a method to detect social bot. First, the user-published text is used as the input to use sentiment analysis and data cleaning to predict the content. Processing and classification, and then adding the quantum similarity algorithm in the emerging quantum discipline in recent years to the structural system of social bot detection to perform similarity clustering, which greatly improves the operating efficiency of the system, and then obtains the corresponding user attribute characteristics for artificial intelligence. Intelligent algorithm classification, on the collected real dataset, marked two social bot groups for machine learning classification and detection, and listed two traditional similarity algorithms for comparison, the results show that the quantum similarity results Compared with the two traditional similarity algorithms, it has improved, and the average accuracy improvement was about 2%.

show abstract

Section: Machine Learning Based Detectionmentioning

confidence: 99%

Section: Work Processmentioning

confidence: 99%

Section: Figurementioning

confidence: 99%

See 1 more Smart Citation

Quantum-based Detection of Higly Semantically Similar Social bot

Liu

2023

FCIS

View full text Add to dashboard Cite

show abstract

“…A fundamental part of the bot detection pipeline corresponds to the computation of features based on Twitter data, and thus a plethora of different types of features have been proposed. Various features are based on content (Ahmed and Abulaish 2013;Gilani, Kochmar, and Crowcroft 2017;Lee, Caverlee, and Webb 2010;Davis et al 2016;Varol et al 2017), sentiment (Loyola-González et al 2019;Dickerson, Kagan, and Subrahmanian 2014;Ferrara et al 2016;Loyola-González et al 2019), account information (Wald et al 2013;Chu et al 2012;Davis et al 2016;Lee, Caverlee, and Webb 2010;Loyola-González et al 2019), usage (Chu et al 2012) and network characteristics (Feng et al 2020;Keller et al 2017;Cresci et al 2017).…”

Section: Introductionmentioning

confidence: 99%

Identification of Twitter Bots Based on an Explainable Machine Learning Framework: The US 2020 Elections Case Study

Shevtsov

Tzagkarakis

Antonakaki

et al. 2022

ICWSM

View full text Add to dashboard Cite

Twitter is one of the most popular social networks attracting millions of users, while a considerable proportion of online discourse is captured. It provides a simple usage framework with short messages and an efficient application programming interface (API) enabling the research community to study and analyze several aspects of this social network. However, the Twitter usage simplicity can lead to malicious handling by various bots. The malicious handling phenomenon expands in online discourse, especially during the electoral periods, where except the legitimate bots used for dissemination and communication purposes, the goal is to manipulate the public opinion and the electorate towards a certain direction, specific ideology, or political party. This paper focuses on the design of a novel system for identifying Twitter bots based on labeled Twitter data. To this end, a supervised machine learning (ML) framework is adopted using an Extreme Gradient Boosting (XGBoost) algorithm, where the hyper-parameters are tuned via cross-validation. Our study also deploys Shapley Additive Explanations (SHAP) for explaining the ML model predictions by calculating feature importance, using the game theoretic-based Shapley values. Experimental evaluation on distinct Twitter datasets demonstrate the superiority of our approach, in terms of bot detection accuracy, when compared against a recent state-of-the-art Twitter bot detection method.

show abstract

“…Since a fundamental part of the bot detection pipeline corresponds to the computation of features based on Twitter data, a plethora of different types of features have been proposed. Various features are based on content (Ahmed and Abulaish 2013;Gilani, Kochmar, and Crowcroft 2017;Lee, Caverlee, and Webb 2010;Davis et al 2016;Varol et al 2017), sentiment (Loyola-González et al 2019;Dickerson, Kagan, and Subrahmanian 2014;Ferrara et al 2016;Loyola-González et al 2019), account information (Wald et al 2013;Chu et al 2012;Davis et al 2016;Lee, Caverlee, and Webb 2010;Loyola-González et al 2019), usage (Chu et al 2012) and network characteristics (Feng et al 2020;Keller et al 2017;Cresci et al 2017).…”

Section: Introductionmentioning

confidence: 99%

Identification of Twitter Bots Based on an Explainable Machine Learning Framework: The US 2020 Elections Case Study

Shevtsov¹,

Tzagkarakis²,

Antonakaki³

et al. 2021

Preprint

View full text Add to dashboard Cite

Twitter is one of the most popular social networks attracting millions of users, while a considerable proportion of online discourse is captured. It provides a simple usage framework with short messages and an efficient application programming interface (API) enabling the research community to study and analyze several aspects of this social network. However, the Twitter usage simplicity can lead to malicious handling by various bots. The malicious handling phenomenon expands in online discourse, especially during the electoral periods, where except the legitimate bots used for dissemination and communication purposes, the goal is to manipulate the public opinion and the electorate towards a certain direction, specific ideology, or political party. This paper focuses on the design of a novel system for identifying Twitter bots based on labeled Twitter data. To this end, a supervised machine learning (ML) framework is adopted using an Extreme Gradient Boosting (XGBoost) algorithm, where the hyper-parameters are tuned via crossvalidation. Our study also deploys Shapley Additive Explanations (SHAP) for explaining the ML model predictions by calculating feature importance, using the game theoreticbased Shapley values. Experimental evaluation on distinct Twitter datasets demonstrate the superiority of our approach, in terms of bot detection accuracy, when compared against a recent state-of-the-art Twitter bot detection method.

show abstract

Towards Learning-Based, Content-Agnostic Detection of Social Bot Traffic

Cited by 9 publications

References 45 publications

Quantum-based Detection of Higly Semantically Similar Social bot

Quantum-based Detection of Higly Semantically Similar Social bot

Identification of Twitter Bots Based on an Explainable Machine Learning Framework: The US 2020 Elections Case Study

Identification of Twitter Bots Based on an Explainable Machine Learning Framework: The US 2020 Elections Case Study

Contact Info

Product

Resources

About