Scalable Cloud-Based Analysis Framework for Medical Big-Data

Abstract

Medical care can be improved by efficient analysis of the large amounts of data involved in patient care. Two important challenges for patient care big-data analysis are the need to cope with unstructured images and text (as well as structured data) and the need for efficient methods that scale for large amounts of data generated by the many patients. The framework described here is designed to cope with these challenges. Firstly it is composed of generic components where the initial stage can transform unstructured image or text data into a set of features for the latter stages. Secondly the machine learning analytics engine incorporates three optimizations which allow the framework to scale and deal with large amounts of unstructured data. The framework is evaluated using both image and text data, and shown to provide an efficient scalable solution.