IST 402 Fraud Informatics The Pennsylvania State University

IST 402: Fraud Informatics The Pennsylvania State University Dongwon Lee, Ph. D. dongwon@psu. edu February 21, 2018 @ IST Teaching Luncheon

Summary of IST 402 l Offered in Fall 2017 l l l 30+ students 2 projects, 1 presentation, 9 labs, 2 reading assignments, 1 final exam Covered major topics 1. 2. 3. 4. 5. 6. Review: Traditional Frauds and basics of ML Fake News Junk Conferences and Predatory Journals Astroturfing, Sockpuppets, and Catfish Spam, Phishing, and Social Engineering Attack Fraud LIKEs 2

1. Review l Traditional Frauds l l l l Credit card fraud Insurance fraud Product warranty fraud Healthcare fraud Telecommunications fraud Money laundering Identity theft Plagiarism 3

1. Review l l Detection vs. Prevention Detection l l l Data-Driven l l l Expert-based Data-Driven Unsupervised Supervised Machine Learning basics l l l Weka Classical models: eg, regression, SVM, ensemble Evaluation 4

2. Fake News l Materials l l News articles: eg, “Inside a fake news sausage factory: This is all about income, ” NY Times, 2016 Research articles: eg, “The spreading of misinformation online, ” PNAS, 2015 Tools: eg, Claim. Buster, B. S. Detector Assignments l l Lab: For news from snopes. com, identify 2 reputable sources (using Wayback machine) Proj #1 (team): Clickbait detection o To build an ML model using Weka to classify if a given headline is clickbait or not 5

6

7

3. Junk Conf. / Predatory Journals l Materials l l Episodes: eg, MIT Prank & SCIgen Research articles: eg, “A physicist experiments with cultural studies, ” Lingua Franca, 1996, Penn State study, CACM 2009 Impact Factor, Bibliometrics Assignments l l Lab #1 Lab #2 8

9 l l l What is the main claim or finding of the article? Is the claim scientifically sound? What do other scientists say about this claim? Is the "Journal of Translational" a questionable journal or not? Argue with evidence.

4. Astroturfing l Materials l l l News articles: eg, “The Best Book Reviews Money Can Buy, ” NY Times, 2012; “The Weird Reasons Why People Make Up False Identities on the Internet, ” Wired, 2014 Research articles: eg, “An Army of Me: Sockpuppets in Online Discussion Communities, ” WWW, 2017 Assignments l Proj #2: Design of fraud detection system 10

11

12 l Using fakespot. com or reviewmeta. com, find/analyze a State College restaurant with questionable reviews

5. Social Engineering Attack l Lab: To build a toy phishing classifier 13

6. Fraud LIKEs l Lab: honeypot experiment, $10 per student 14

15 Data Collection Broker-Initiated Market Fake Likers Legit Likers Buyer-Initiated Market

Honeypot Page 16

Takeaways l Students find topics familiar and comfortable l l l Abundance of learning materials l l l News articles, You. Tube videos, Social Media Eg Games and tools Challenges l l l Eg, fake reviews, craiglist scams, fake news Well engaged in class activities and discussion Varying computational/ML skills Understanding research articles Plan l IST 402 (2017 F) IST 497 (2018 F) DS 4 xx 17