Klasifikace typologie metabolismu na základě analýzy dat z energometrickýzch testů

Abstract

This work deals with the classification of metabolism typology based on analysis of the energometry tests. Three data sources from patient are used in this work. The sources mean data from respiratory quotient measurement and calculated indicators of food utilization. Further the data from biochemical analysis of blood and insulin test are also used in analysis. At last, the measurement of bio-impedance and patient description are used in the work. The patients are divided into two groups, analytical and test group. The analysis group is processed by principal component analysis and it is normalized. The work also includes methods for fuzzy classification and estimation of classification parameters. After the classification space estimation the algorithm is tested by the testing group and evaluated by statistical hypothesis testing. The aim of the thesis is design algorithm for an automatic diagnostic system which insert patient to the group of features based on the previous model.