The Exponentiated Power Shanker Distribution With Application L. as Potential Sources of Xanthine Oxidase and α-Glucosidase Inhibitors
N. Badmus, A. de Souza, and O. Faweya (1-14)
Abstract
A new exponentiated power Shanker distribution is proposed, and its statistical properties are thoroughly discussed. The distribution parameters are estimated using the maximum likelihood estimation method. The new distribution is then extended to a regression model by applying the logarithmic transformation. The proposed regression model is applied to a marriage dissolution data set consisting of 568 entries, where the response variable is the number of years in the marriage and the predictor variables are the husband’s education level, husband’s race (black or not), marital mixing (mixed or not), and divorce status. The results reveal that some predictors significantly influence the duration of marriage. The estimated regression model is given by year(xˆ) = 0.0809 − 1.6190(Heduc) + 0.4343(Heblack) − 1.0747(Mixed) + 1.1819(Divorce). The model selection criteria for the propose model yield the following values: Akaike information criterion = 8073.054, Bayesian information criterion = 8086.080, consistent Akaike information criterion = 8087.080, and Hannan-Quinn information criterion = 8078.137. Furthermore, a machine learning-based multiple regression model is conducted to estimate mean absolute error, mean squared error, and root mean squared error, which are 10.1138, 162.1605, and 12.7342, respectively. The results are compared with those obtained from a traditional linear regression model, and findings suggest that the machine learning-based multiple regression model outperforms the linear regression model, with smaller error values.