Sina-weibo spammer detection with GBDT

Yang Qiao*, Huaping Zhang, Min Yu, Yu Zhang

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

3 Citations (Scopus)

Abstract

In China, Sina-Weibo, with its rising popularity as a microblogging website, has inevitably attracted the attention of spammers. Spammers use myriad of techniques to evade security mechanisms and post spam messages, which are either unwelcome advertisements for the victim or lure victims in to clicking malicious URLs embedded in spam tweets. With the extensive application of machine learning in social media mining and Sina-Weibo’s development, we get many new ideas for the spammers detection. In this paper, we first make a comprehensive analysis specifically aiming at some new Sina-Weibo features rather than other social media, we further design a new feature set to detect spammers. We grab a large amount of Sina-Weibo data on the Internet and train the classifier with the algorithm GBDT. Through our experiments, we show that our new designed features are much more effective than some existing detector. And GBDT also has been significantly improved in both the accuracy and the FP-rate.

Original languageEnglish
Title of host publicationSocial Media Processing - 5th National Conference, SMP 2016, Proceedings
EditorsHongfei Lin, Yuming Li, Guoxiong Xiang, Mingwen Wang
PublisherSpringer Verlag
Pages220-232
Number of pages13
ISBN (Print)9789811029929
DOIs
Publication statusPublished - 2016
Event5th National Conference on Social Media Processing, SMP 2016 - Nanchang, China
Duration: 29 Oct 201630 Oct 2016

Publication series

NameCommunications in Computer and Information Science
Volume669
ISSN (Print)1865-0929

Conference

Conference5th National Conference on Social Media Processing, SMP 2016
Country/TerritoryChina
CityNanchang
Period29/10/1630/10/16

Keywords

  • Detection
  • GBDT
  • Social media
  • Spammer

Fingerprint

Dive into the research topics of 'Sina-weibo spammer detection with GBDT'. Together they form a unique fingerprint.

Cite this