Using Normalised Compression Distance to Identify Different Profiling Days in Type 1 Diabetic Patients

Abstract

This work is devoted to providing patients and physicians with a novel tool to analyse and extract information for better management of type I diabetes. We use a clustering methodology based on the normalised compression distance to identify different profiles of days. The methodology has been validated using data generated by a simulator of virtual patients, which include an exercise model. Profiling daily data can help physicians and patients cope with information overload and assist in future planning for improved treatments and self-management of diabetes type 1