Using Logistic Regression and Discriminant Analysis Techniques for Factors Affecting Heart Diseases Infection (A comparative study of Cardiac surgery and renal transplant Center at Ahmed Gasi

Abstract

Diagnosing cardiac diseases by using an advanced statistical classifying model is vitally important for cardiac diseases which are viewed as serious diseases which can put an end to a patient’s life, therefore, early diagnosis should be done through classification and to differentiate among the cardiac patients and non-cardiac ones based on the factors that can cause the cardiac diseases by using technologies such as logistic regression and discriminative analysis. This study aimed at identifying the main factors that lead to the cardiac diseases and to investigate the effect of each one of these factors and to arrange them in terms of the preference as in the model and its ability to differentiate and classify. A statistical model was constructed so as to predict the possibility of infection using a binary logistic regression as well as discriminatory analysis. The descriptive analytical method was adopted to describe the factors affecting the infection, the research data were collected from the Cardiac Surgery Centre and renal transplant, a study sample was taken from 214 the infected patients, and a controlled sample of 214 from non- infected patients which were specified by using Stephan Thompson Equation, in addition to the analytical method represented by the binary logistic regression and discriminatory analytical method for building the best model through SPSS programme. This study arrived at the most important findings: the amount of the binary logistic regression that include the affecting factors (hypertension, gender, diabetes, calestrol, heritage, and weight) has high efficiency for classification and prediction about 91.8%, the rate of infection of these factors account for 82.3%, hypertension 54.5%, gender accounts for 15.7%, diabetes 14.55, calestrol 8.3% heritage 4.4% and weight accounts for 4.1%. The estimated discriminatory model includes effective factors; hypertension, diabetes, heritage, calestrol, weight, and gender have high discriminatory that accounts for 91,1% and the contributing rate of these factors 72.4% and hypertension contributes greatly in terms of discrimination that is 82,1%, heritage 45.3%, diabetes 44.8%, calestrol 40,6%, weight 21.1%, and gender 18.8%. the ability of the binary logistic regression model is better than the discriminatory model because of It has the highest efficiency and the ability to diagnose and predict the possibility of infection with minimum error at 8.2.The study recommended that further studies should be conducted to investigate other affecting factors such as age, smoking, and sports, to investigate their impact on cardiac patients in comparison to this study.