Miroslava Nedyalkova
Sofia University, Bulgaria
Title: Assessment of similarity patterns of diabetes mellitus type 2 patient's clinical data by chemometric tools
Biography
Biography: Miroslava Nedyalkova
Abstract
The present study deals with the assessment and interpretation of clinical data for patients diagnosed with diabetes mellitus type 2 by the use of intelligent data analysis. One hundred (100) patients were involved and each of them were characterized by 33 clinical indicators related to testing the blood sugar, renal and liver functions and anthropometric data. The assessment methods were non-hierarchical cluster analysis being very convenient for supervised pattern recognition tasks. The major goal of the study is to try to find patterns of similarity (clusters) within the data set in order to improve the overall information needed by the medics for specific treatment and checking of the health status. Five clusters were interpreted with respect to similarities between the clinical indicators. The same number of clusters was a priori chosen for interpretation of the patterns of similarity between the patients. Each pattern will be specifically explained and discussed.