A Modified KL Estimator for the Binary Logistic Regression Model

• Published in: 2023 International Conference on Science, Engineering and Business for Sustainable Development Goals (SEB-SDG) Vol. 1; pp. 1 - 5
• Main Authors: Aladeitan, Benedicta B., Adebimpe, Olukayode, Ayinde, Kayode, Lukman, Adewale, Oludoun, Olajumoke, Abiodun, Elizabeth
• Format: Conference Proceeding
• Language: English
• Published: IEEE 05-04-2023
• Subjects: Business; Data models; KL estimator; Linear regression; Logistic Regression; Logistics; Maximum likelihood estimation; Mean square error methods; Modified Kibria Lukman; Multicollinearity; Sustainable development
• DOI: 10.1109/SEB-SDG57117.2023.10124573

The logistic regression model is used in place of the linear regression model when the dependent variable is primarily dichotomous. Multicollinearity occurs when the independent variables in a logistic regression model are correlated, particularly if the independent variables are continuous. The maximum likelihood estimation procedure is the process for estimating the parameters for the logistic regression model. The Maximum Likelihood Estimator performs poorly in the presence of multicollinearity therefore alternative techniques for estimating the parameters of the logistic regression model are being sought out. Therefore, the goal of this study is to create a new estimator for the logistic regression model under multicollinearity. The efficiency of the new estimator is tested using a simulation and a real-world study, both of which use the mean square error as a criterion. The simulation and real-world study's findings indicate that the new estimator outperform the logistic regression model's other one-parameter estimators.