ISIS OEC members for classification It is worth noting that our classifier did

Isis oec members for classification it is worth

This preview shows page 19 - 20 out of 20 pages.

ISIS OEC members for classification. It is worth noting that our classifier did not simply find accounts contained in clusters 4 and 6 as is highlighted by the figure as well. We constructed a feature set using spectral representations of the F rec , M rec , and H user × user ; sharedHashTag networks as described in Section 2. A full list and description of our feature set is included in Table 2 . As will be highlighted in Section 4, the ISIS OEC is highly interested in the ongoing operations in Northern Iraq and Syria. As such, they discuss political figures and news sources extensively. Initial attempts to detect the ISIS OEC contained many official accounts as previously defined. Therefore, in our first iteration of multiplex vertex classification (MVC) the task was to remove all official, celebrity, and news media accounts. To do so, we conduct an iteration of IVCC by developing a training set of positive and negative examples of official accounts to apply to the rest of our dataset. Our positive case labels for official accounts consisted of 2,144 known celebrities, politicians, and journalists as well as an additional 873 accounts with more than 150,000 followers. We labelled the 8,356 suspended/deleted accounts in our dataset as non-official accounts, and trained a Random Forest classifier [ 55 ] The Random Forest classifier is an ensemble method that constructs a multitude of decision trees and uses the mode of these classes to correct for the problem of overfitting associated with many tree based classifiers. We found its performance to be significantly better than SVM with respect to accuracy when identifying official accounts to remove from our dataset. The classifier’s superior performance was likely due to the various types of official accounts creating contingencies better captured by a tree based classifier. It is worth mentioning that we are not interested in using this classifier on accounts not contained in G ; so we conduct use a train/test split, but also use random sampling to assess accuracy. Conclusion The present work makes two major contributions to the literature. First, we develop iterative vertex clustering and classification (IVCC), a scalable, annotated network analytic approach for extremist community detection in social media. Our approach outperforms two existing approaches on a classification task of identifying ISIS supporting users by a significant margin. Second, we provided an illustrative case study of the ISIS supporting network on Twitter. To the best of our knowledge, it is the
Image of page 19
most comprehensive study of this network, and it provides a variety of important insights that may prove important in better understanding the incredible proliferation of ISIS propaganda on Twitter.
Image of page 20

You've reached the end of your free preview.

Want to read all 20 pages?

  • Fall '17

What students are saying

  • Left Quote Icon

    As a current student on this bumpy collegiate pathway, I stumbled upon Course Hero, where I can find study resources for nearly all my courses, get online help from tutors 24/7, and even share my old projects, papers, and lecture notes with other students.

    Student Picture

    Kiran Temple University Fox School of Business ‘17, Course Hero Intern

  • Left Quote Icon

    I cannot even describe how much Course Hero helped me this summer. It’s truly become something I can always rely on and help me. In the end, I was not only able to survive summer classes, but I was able to thrive thanks to Course Hero.

    Student Picture

    Dana University of Pennsylvania ‘17, Course Hero Intern

  • Left Quote Icon

    The ability to access any university’s resources through Course Hero proved invaluable in my case. I was behind on Tulane coursework and actually used UCLA’s materials to help me move forward and get everything together on time.

    Student Picture

    Jill Tulane University ‘16, Course Hero Intern

Ask Expert Tutors You can ask You can ask ( soon) You can ask (will expire )
Answers in as fast as 15 minutes