HEALTHCARE TEXT ANALYTICS USING RECENT ML TECHNIQUES AND DATA CLASSIFICATION USING AWS CLOUD ML SERVICES

Abstract

Classification of clinical texts has a significant impact on disease diagnosis, medical research and automated development of disease ontologies. Because they contain terms that describe medical concepts and terminology, the data set is quite noisy and the text in the transcriptions overlaps with the categories making clinical text difficult to classify. The clinical narrative, which provides a patient’s history and evaluations as well as data for clinical decision-making, is the main form of communication in the medical field. The aim of the study is to make disease diagnoses based on medical records using ML algorithms. The proposed clinical text classification model using weak monitoring to reduce the human efforts to create labeled training data and conduct feature engineering. The primary objective is to contrast this approach with a logistic regression model to classify medical records clinical text and expect superior performance compared to the logistic regression model for an imbalanced medical transcriptions dataset. A promising intelligent data-driven health system to archive and classify healthcare records relies on the ability to extract and contextualize unstructured medical data in a form of a single easy-to-use API by leveraging Machine Learning (ML) services from Amazon cloud in a clinical workflow. AWS services such as S3, Textract, Comprehend Medical, and DynamoDB can be integrated to create a comprehensive solution for handling medical document processing, extracting medical information, performing medical text analysis, and storing the data in a structured manner.