Analiza utjecaja promatranih varijabli na bolest srca logističkom regresijom

master's thesis

• Title (english): Analysis of the influence of the observed variables on heart disease by logistic regression
• Author: Andrea Barbarić
• Mentor: Anamarija Jazbec (mentor)
• Committee member: Anamarija Jazbec (predsjednik povjerenstva)
• Committee member: Vanja Wagner (član povjerenstva)
• Committee member: Marko Tadić (član povjerenstva)
• Committee member: Ljiljana Arambašić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2021-03-05, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

U ovome radu opisane su glavne značajke logističkog regresijskog modela te njegova usporedba s linearnom regresijom. Predstavljen je način procjene i interpretacije parametara u modelu te odabir najadekvatnijeg modela. Zatim je u drugom dijelu ta teoretska podloga primijenjena na konkretnim podacima. Točnije, baza podataka sastoji se od 297 opservacija koje predstavljaju pacijente i 14 varijabli koje predstavljaju određene podatke o njihovu zdravstvenom stanju na prijemu u bolnicu. Na temelju toga, cilj je bio procijeniti koje varijable i koliko statistički značajno utječu na pojavu bolesti srca kod osobe. Pri izradi najadekvatnijeg modela korištene su univarijatna i multivarijatna logistička regresija te procedura Stepwise. Dobiveno je da na razvoj bolesti srca najviše utječu spol, tip boli u prsima, krvni tlak u mirovanju, vršni nagib ST-segmenta u pokretu, broj vidljivih glavnih krvnih žila tijekom fluoroskopije, rezultati talij testa, te angina izazvana tjelesnom aktivnošću.

Abstract (english)

This thesis describes the main features of the logistic regression model and its comparison with linear regression. It presents the method of estimation and interpretation of model parameters as well as the selection of the most adequate model. In the second part, this theoretical basis was applied to concrete data. Specifically, the database consists of 297 observations representing patients and 14 variables representing information about their health status at hospital admission. Based on this, the aim was to assess which variables have statistically significant effect on the occurrence of heart disease. Univariate and multivariate logistic regression and Stepwise procedure were used in the development of the most adequate model. The results have shown that heart disease is mostly affected by gender, chest pain type, resting blood pressure, the slope of the peak exercise ST-segment, number of major vessels colored by fluoroscopy, thallium test results and exercise induced angina.

• Keywords: logistički regresijski model; bolesti srca; procedura Stepwise
• Keywords (english): logistic regression model; heart disease; Stepwise procedure
• Language: croatian
• URN:NBN: urn:nbn:hr:217:957320
• Study programme: Title: Mathematical Statistics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra matematike (magistar/magistra matematike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access; Embargo expiration date: 2021-12-31
• Created on: 2021-03-23 09:42:44