A Modified Half-Logistic Distribution With Regression Analysis

This study introduces the power odd Lindley half-logistic distribution (POLiHLD), a novel statistical distribution developed to provide enhanced flexibility for modeling diverse data sets. This distribution is derived by combining the characteristics of the odd Lindley and half-logistic distributions through the power transformation, resulting in a model capable of capturing various shapes and tail behaviors. We explore the fundamental statistical properties of the POLiHLD, including moments and moment-generating functions, a sequential probability ratio test, and average sample number, among others. Extensive simulation studies were conducted to validate the estimation method used to estimate the parameters of the developed distribution. These simulations highlight the robustness of the estimation method used. The POLiHLD's flexibility is shown through its fitting to two real datasets: the first data set represents the ordered failure of components, and the second data set captures economic data. Comparative analyses with existing distributions show that the POLiHLD provides a better fit for the analyzed datasets. A regression model was developed to ascertain the predictive ability of the proposed model.