A corpus for mining drug-related knowledge from Twitter chatter: Language models and their utilities

Published: 17 July 2017| Version 3 | DOI: 10.17632/dwr4xn8kcv.3
Abeed Sarker,
Graciela Gonzalez-Hernandez


Language models. As described in the publication titled above. DSM-langauge-models-3M-LARGE is generated from over 3M posts using window size 5 and dimension 400. **USE THIS**: DSM-language-model-1B-LARGE is generated from ~ 1B tweets from user timelines where at least 1 medication is mentioned. This model is an n-gram model.