We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:


References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computers and Society

Title: Detection and Characterization of Illegal Marketing and Promotion of Prescription Drugs on Twitter

Abstract: Illicit online pharmacies allow the purchase of prescription drugs online without a prescription. Such pharmacies leverage social media platforms such as Twit- ter as a promotion and marketing tool with the intent of reaching out to a larger, potentially younger demographics of the population. Given the serious negative health effects that arise from abusing such drugs, it is important to identify the relevant content on social media and exterminate their presence as quickly as pos- sible. In response, we collected all the tweets that contained the names of certain preselected controlled substances over a period of 5 months. We found that an unsupervised topic modeling based methodology is able to identify tweets that promote and market controlled substances with high precision. We also study the meta-data characteristics of such tweets and the users who post them and find that they have several distinguishing characteristics that sets them apart. We were able to train supervised methods and achieve high performance in detecting such content and the users who post them.
Comments: NIPS 2017 Workshop Machine Learning for Health
Subjects: Computers and Society (cs.CY)
Cite as: arXiv:1712.00507 [cs.CY]
  (or arXiv:1712.00507v1 [cs.CY] for this version)

Submission history

From: Janani Kalyanam [view email]
[v1] Fri, 1 Dec 2017 22:34:07 GMT (12kb)

Link back to: arXiv, form interface, contact.