Distant Supervision with Transductive Learning for Adverse Drug Reaction Identification from Electronic Medical Records

Figure 1

An example of narrative notes from a discharge summary in an EMR system is shown in (a). The possible outcomes derived by NERs, normalization, and relation generation of drugs and events from the given texts are displayed in (b). Both drugs and events are unified by UMLS CUI. For privacy concerns, confidential information is concealed using deidentification as [].