Performance analysis of deep neural models for automatic identification of disease status

Abstract

An early detection of the disease status can play a vital role, as timely action can save patient's life. A disease status of the patient is a type of PHI (Protected Health Infonnation).In general the process ofidentifying the PHI is done manually on the structured dataset, this activity is prohibitively expensive, time consuming and error prone.In order to overcome the known drawback of the manual identification system an implementation of automated personal health information identification like disease status is required. In our work an effort has been done to build a model with a novel approach based on deep neural models for automatic identification of disease status (one of the PHIs) from free text clinical records. The experimental evaluation of the proposed models on two different datasets from i2b2 Challenge tasks, show the merit ofthe proposed scheme.