Using a binary logistic regression model to diagnose the effect of factors causing cancer : An applied study on a sample of patients at the Oncology Hospital in Baghdad

Abstract

Since the data about this phenomenon is binary, meaning the presence or absence of an event, so it applied the binary logistic regression model for the binary response variable in which the dependent variable is either equal to one for the occurrence of the event or zero for the absence of the event. The research aims to use this model to obtain the most important factors affecting cancerous tumors. The study included cancerous tumor data from the Ministry of Health and the Tumor Teaching Hospital, The statistical program (SPSS). The conclusions showed that the logistic regression model is suitable for testing the data and does not suffer from the problem of multilinearity. As well as six factors affecting cancer diseases (Smoking, chronic diseases, family history, weight, height, and marital status), but sex does not have a significant effect on cancerous diseases, according to the sample drawn.

Subject Classification:

• Primary 93A30

Keywords:

• Cancer
• Logistic regression model
• Maximum likelihood
• Quality tests
• Hosmer-lemeshow test
• Coefficient of determination