Multicollinearity in Logistic Regression Model -Subject Review-

Abstract

The logistic regression model is one of the modern statistical methods developed to predict the set of quantitative variables (nominal or monotonous), and it is considered as an alternative test for the simple and multiple linear regression equation as well as it is subject to the model concepts in terms of the possibility of testing the effect of the overall pattern of the group of independent variables on the dependent variable and in terms of its use For concepts of standard matching criteria, and in some cases there is a correlation between the explanatory variables which leads to contrast variation and this problem is called the problem of Multicollinearity. This research included an article review to estimate the parameters of the logistic regression model in several biased ways to reduce the problem of multicollinearity between the variables. These methods were compared through the use of the mean square error (MSE) standard. The methods presented in the research have been applied to Monte Carlo simulation data to evaluate the performance of the methods and compare them, as well as the application to real data and the simulation results and the real application that the logistic ridge estimator is the best of other method. 1-Introduction Regression is a statistical method that specializes in studying the relationship between a dependent variable and one or several other independent variables, resulting in a mathematical equation where this relationship represents the best representation. The logistic regression model is a special case of the generalized linear model which is the most common in analyzing metadata and is a logarithmic transformation of linear regression, and it has several types, but the most common is the analysis of the binary logistic regression that we will use in our research without other types of logistic regression. it is a more powerful tool because it provides a test of the significance of parameters, and it also gives the researcher an idea of how much the independent variable affects the qualitative dependent variable dual value In addition, it sees the effect of independent variables, which allows the researcher to conclude that a variable is considered stronger than the other variable in understanding the appearance of the desired result, and that the logistic regression analysis can include qualitative independent variables The effect of the interaction between the independent variables in the two-valued dependent variable [Abbas,2012]. The researcher faces many problems, most of which are the lack of analysis hypotheses when using the method of ordinary least squares, including the problem of multicollinearity that affects the results of estimates and tests, and this problem appears as a result of an association between explanatory variables that lead to giving weak estimates that cannot be relied upon as the variations of these The capabilities are amplified and unacceptable and the (OLS) method is not able to give good estimates when there is a linear relationship between the explanatory variables. 2-Logistic Regression The logistic regression model is an important statistical model in analyzing binary data (0 or 1) as the primary goal of most studies is to analyze and evaluate relationships between a set of variables to obtain a