SCAN: A shared causal attention network for adverse drug reactions detection in tweets

Twitter is a popular social media site on which people post millions of Tweets every day. As patients often share their experiences with drugs Twitter, can also be considered as rich alternative source adverse drug reaction (ADR)-related information. This information useful for health authorities and manufacturing companies to monitor the post-marketing effectiveness drugs. However, automatic detection ADRs in challenging, are informal prone grammatical errors. The existing approaches automatically detecting do not consider cause-effect relationships between an ADR. In this paper, we propose novel shared causal attention network that exploits such detect Tweets. our approach, split Tweet into prefix, midfix, postfix segments based position name separately extract features from segments. We then these separate both word parts-of-speech features, apply multi-head self-attention mechanism. run extensive experiments three publicly available benchmark datasets illustrate proposed approach.