Named Entity Recognition (NER) autodetection and Adverse Events (AEs) prediction from Social Media and scientific journals using a Deep Learning approach

Social medias data bases are important for continuous and automated Adverse Drug Reactions (ADRs) surveillance. Predicting ADRs can reduce the related mortality. A systematic review of the medical scientific literature is required for tracking and identifying the risk/benefit ratio of drugs and safety issues.
The implementation of means of standardizing patient’s language used in social media such as Twitter will improve the precision of ADR detection and continuous surveillance at post-marketing phase of drugs (off-label drug uses included). The required processing flow will be improved with labeling mechanisms enhancing machine learning algorithms that can make them more accurate.

Faculty Supervisor:

Sylvie PERREAULT;Chahé Nerguizian

Student:

Partner:

Medvalgo Group

Discipline:

Life Sciences

Sector:

Professional, scientific and technical services

University:

Université de Montréal

Program: