Mining patterns in Big Data K-H networks

Abstract

Can keyword-hashtag networks, derived from Big Data environments such as Twitter, yield clinicians a powerful tool to extrapolate patterns that may lead to development of new medical therapy and/or drugs? In our paper, we present a systematic network mining method to answer this question. We present HashnetMiner, a new pattern detection algorithm that operates on networks of medical concepts and hashtags. Concepts are selected from widely accessible databases (e.g., Medical Subject Heading [MeSH] descriptors, and Drugs.com), and hashtags are harvested from the associations with concepts that appear in tweets. The algorithm discerns promising discoveries that will be further explained in this paper. To the best of our knowledge, this is the first effort that uses Big Data networks mining to address such a question.