Modeliranje dijabetesa pomoću logističke regresije s nominalnom zavisnom varijablom

Dodik, Anto (2015) Modeliranje dijabetesa pomoću logističke regresije s nominalnom zavisnom varijablom. Diploma thesis, Faculty of Science > Department of Mathematics.

Language: Croatian

Download (617kB) | Preview


Logistic regression, a powerful statistic tool, is used in researches which include categorical dependent variable. Logistic regression is robust thanks to the fact that it requires a minimum set of assumptions. As a result, it can be used widely, in scientific as well as in practical context. This thesis describes the basic concepts of logistic regression. Among others, the terms of odds ratio, maximum likelihood estimation and logistic regression model have been analysed. Logistic regression model is used to predict the probability of the event occurring by adjusting the data to the logistic curve. Theoretical background of logistic regression with dichotomous dependent variable is implemented in modeling of type 1 and type 2 diabetes. The database from the course Statistical methods in Biomedical Research has been examined. Based on univariate and multivariate logistic regression, as well as on Stepwise procedure in SAS, the most adequate model for the prediction of type 1 and type 2 diabetes has been found. The results have shown that hemoglobin and diastolic blood pressure are the main predictors of type 1 diabetes, while the hemoglobin and glucose are predictors for type 2 diabetes.

Item Type: Thesis (Diploma thesis)
Supervisor: Jazbec, Anamarija
Date: 2015
Number of Pages: 43
Subjects: NATURAL SCIENCES > Mathematics
Divisions: Faculty of Science > Department of Mathematics
Depositing User: Iva Prah
Date Deposited: 07 Oct 2015 09:42
Last Modified: 07 Oct 2015 09:42

Actions (login required)

View Item View Item

Nema podataka za dohvacanje citata