Inconsistency-Detection-in-Medical-Annotations
It's one of the problems given by ezDI Inc. during Smart India Hackathon 2019.
Problem Statemnt:
Text in clinical documents is annotated to tag the words or phrases as problems, diagnoses, procedures, drugs etc and identify the relationships between them. Semantically similar words can be annotated differently by different annotators due to ambiguity in understanding the guidelines,incomplete experience or human errors. Identifying inconsistencies can help in improving the data quality You need to develop a tool to display these similar entities which have different patterns of annotations.
Solution Abstract:
We propose a dynamic system for clustering the similar entities and toassociate the annotation patterns which will be updated in a databasemaintained using MySQL which will be fed to a web application hostedon Google Cloud which will display the result in a fast, scalable and reliable manner. The detailed architecture of the system is shown below.
