Model predikcije malignosti tumora dojke logističkom regresijom

master's thesis

• Author: Jelena Krajina
• Mentor: Anamarija Jazbec (mentor)
• Committee member: Anamarija Jazbec (predsjednik povjerenstva)
• Committee member: Miljenko Huzak (član povjerenstva)
• Committee member: Goran Muić (član povjerenstva)
• Committee member: Marko Vrdoljak (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2017-09-27, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

U ovom radu napravili smo model predikcije malignosti tumora dojke. Tumor dojke je zloćudna bolest koja napadna sve više i više žena, ali i muškaraca. Oko trećinu svih malignih tumora u žena čini upravo tumor dojke. Također tumor dojke je visoko zastupljen među uzrocima smrti žena. U radu smo koristili podatke „Breast Cancer Wisconsin (Diagnostic) Data Set“ koje je prikupio dr. William H. Wolber (University of Wisconsin Hospitals, Madison). Podaci sadrže 699 opservacija i 11 varijabli, odnosno atributa. Pri obradi podataka koristili smo metodu logističke regresije i statistički program SAS. Nakon provedenih analiza došli smo do zaključka da su varijable: Debljina grumena, Marginalna adhezija, Gole jezgre, Normalan kromatin i Normalne jezgrice statistički značajne za model predikcije malignosti tumora dojke.

Abstract (english)

In this paper, we have made a model for prediction of breast tumor malignancy. Breast cancer is a malignant disease that attacks more and more women, but also men. About a third of all malignant tumors in women are breast tumors. Also, breast cancer is highly represented among the causes of the death of a woman. In this paper we used the data ”Breast Cancer Wisconsin (Diagnostic) Data Set” collected by Dr. William H. Wolber (University of Wisconsin Hospitals, Madison). The data contains 699 observations and 11 variables, respectively attributes. When processing the data we have used the method of logistic regression and the statistical program SAS. After the analysis we have come to the conclusion that the variables : Clump Thickness, Marginal Adhesion, Bare Nuclei, Bland Chromatin and Normal Nucleoli are statistically significant for the model of prediction of breast tumor malignancy.

• Keywords: model predikcije; malignost tumora; tumor dojke; rak dojke; metoda logističke regresije; SAS
• Keywords (english): prediction model; tumor malignancy; breast tumor; breast cancer; logistic regression model; SAS
• Language: croatian
• URN:NBN: urn:nbn:hr:217:112318
• 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
• Created on: 2018-02-01 13:31:03