Quantcast

A twitter recruitment intelligent system: association rule mining for smoking cessation

Research paper by Ahmed Abdeen Hamed, Xindong Wu, Alan Rubin

Indexed on: 12 Aug '14Published on: 12 Aug '14Published in: Social Network Analysis and Mining



Abstract

Digital recruitment is increasingly becoming a popular avenue for identifying human subjects for various studies. The process starts with an online ad that describes the task and explains expectations. As social media has exploded in popularity, efforts are being made to use social media advertisement for various recruitment purposes. There are, however, many unanswered questions about how best to do that. In this paper, we present an innovative Twitter recruitment system for a smoking cessation nicotine patch study. The goals of the paper are to: (1) present the approach we have taken to solve the problem of digital recruitment; (2) provide the system specification and design of a rule-based system; (3) present the algorithms and data mining approaches (classification and association analysis) using Twitter data; and (4) present the promising outcome of the initial version of the system and summarize the results. This is the first effort to introduce a practical solution for digital recruitment campaigns that is large-scale, inexpensive, efficient and reaches out to individuals in near real-time as their needs are expressed. A continuous update on how our system is performing, in real-time, can be viewed at https://twitter.com/TobaccoQuit.