Volume 13 Number 7 (Jul. 2018)
Home > Archive > 2018 > Volume 13 Number 7 (Jul. 2018) >
JCP 2018 Vol.13(7): 823-829 ISSN: 1796-203X
doi: 10.17706/jcp.13.7.823-829

A Hybrid Method for Spammer Detection in Social Networks by Analyzing Graph and User Behavior

Mona Mona Mona Najafi Sarpiri1, Taghi Javdani Gandomani2, Mahsa Teymourzadeh3, Akram Motamedi4
1Comp. Dept., Dolat Abad Branch, Islamic Azad University, Dolat Abad, Isfahan, Iran.
2Comp. Dept., Boroujen Branch, Islamic Azad University, Boroujen, Iran.
3Comp. Dept., Shahrekord Branch, Islamic Azad University, Shahrekord, Iran.
4Comp. Dept., Najaf Abad Branch, Islamic Azad University, Najaf Abad, Iran. ….

Abstract—With the increasing use of the Internet and social networks, there are many spammers causing security problems and numerous challenges in these services. Detection of spammers has attracted much attention in the recent years and several strategies have been proposed for detection and limitation of their activities by different researchers. However, there are still many challenges and open questions in this area which need further research. Although there are still many problems in this area needs further researches. This study proposes a graph analysis based method for spammer detection by analyzing their behaviors and their relation with the users. Finally, a solution is provided to facilitate the detection process. The aim of this paper, by applying the hybrid graph analysis method and behavior analysis, is to increase the diagnostic accuracy and detection rate with the help of appropriate classification algorithms and the most effective features. So, two scenarios were used to achieve higher accuracy level and lower false positive. The first scenario was based on using the entire data to build and evaluate the model. The results showed that despite the high precision of this approach, due to the high levels of false positive, this approach is not appropriate. In the second scenario, the ratio of the normal users to spammers was considered equal to 2 to 1 which led to satisfactory results. After reviewing the confusion matrix and false positives in different algorithms, the Logistic algorithm was chosen as an appropriate algorithm which meets the objective of this study.

Index Terms—Spam, spammers, spam detection, user behavior analysis, graph analysis.


Cite: Mona Najafi Sarpiri, Taghi Javdani Gandomani, Mahsa Teymourzadeh, Akram Motamedi, "A Hybrid Method for Spammer Detection in Social Networks by Analyzing Graph and User Behavior," Journal of Computers vol. 13, no. 7, pp. 823-829, 2018.

General Information

ISSN: 1796-203X
Frequency: Monthly (2006-2014); Bimonthly (Since 2015)
Editor-in-Chief: Prof. Liansheng Tan
Executive Editor: Ms. Nina Lee
Abstracting/ Indexing: DBLP, EBSCO,  ProQuest, INSPEC, ULRICH's Periodicals Directory, WorldCat, CNKI,etc
E-mail: jcp@iap.org
  • Apr 24, 2018 News!

    Vol 13, No 8 has been published with online version   [Click]

  • Dec 26, 2017 News!

    Vol 12, No 1-N0 5 has been indexed by EI (Inspec)     [Click]

  • Dec 26, 2017 News!

    Vol 11, No 4-N0 6 has been indexed by EI (Inspec)     [Click]

  • Dec 21, 2017 News!

    Vol 13, No 7 has been published with online version 12 papers are published in this issue after peer review   [Click]

  • Sep 26, 2017 News!

    Papers published in JCP Volume 12 have all been indexed by DBLP   [Click]

  • Read more>>