Feldman, Ronen, et al. "Utilizing text mining on online medical forums to predict label change due to adverse drug reactions." Proceedings of the 21th ACM SIGKDD international conference on knowledge discovery and data mining. 2015.

This paper introduces an end-to-end text mining methodology for extracting adverse drug reactions (ADRs) from medical forums on the Web. The methodology is distinctive due to three key features: (i) it employs a head-driven phrase structure grammar (HPSG) based parser; (ii) it leverages domain-specific relation patterns, acquired mainly through unsupervised methods applied to a large, unlabeled text corpus; and (iii) it uses automated post-processing algorithms to enhance the set of extracted relations. The approach is empirically validated by demonstrating its ability to predict ADRs before they are reported by the Food and Drug Administration (FDA), effectively indicating ADRs that were not detected in clinical trials but were later reported by the FDA as label changes.