Applying the Binary Logistic Regression Analysis on The Medical Data

Science Journal of University of Zakho

Qais M. Abdulqader¹ 

[1] Duhok Polytechnic University;

关键词: Logistic regression;    Hosmer–Lemeshow test;    Likelihood ratio test;    Maximum likelihood estimation;    Wald test;   

DOI  :  10.25271/2017.5.4.388

来源: DOAJ

【 摘 要 】

In this paper, the Binary Logistic Regression Analysis BLRA technique has been used and applied for building the best model for Hepatitis disease data using best subsets regression and stepwise procedures and depending on some laboratory tests such as glutamate oxalate transaminase, glutamate pyruvate transaminase, alkaline phosphatase, and total serum bilirubin which represents explanatory variables. Also, the technique has used for classifying persons into two groups which are infected and non-infected with viral Hepatitis disease. A random sample size consists of 200 persons has been selected which represents 86 of uninfected and 114 of infected persons. The results of the analysis showed that first, the two procedures identified the same three explanatory variables out of four and they were statistically significant, and it has been reliable in building the logistic model. And second, the percentage of visible correct classification rate was about 98% which represents the high ability of the model for classification.

【 授权许可】

Unknown