广义指数几何分布的扩展及参数估计  

Extension and Parameter Estimation of Generalized Exponential Geometric Distribution

作　　者：陈世彤 田野 宓颖[1] 李树有[1] CHEN Shi-Tong;TIAN Ye;MI Ying;LI Shu-you(College of Science,Liaoning University of Technology,Jinzhou 121001,China)

机构地区：[1]辽宁工业大学理学院,辽宁锦州121001

出　　处：《辽宁工业大学学报（自然科学版）》2024年第3期152-155,163,共5页Journal of Liaoning University of Technology(Natural Science Edition)

基　　金：辽宁省教育厅基础研究项目(JJL202015409)。

摘　　要：在广义指数几何分布的基础上对参数进行扩展,提出了一种新的寿命分布。旨在通过这种扩展,增强模型针对一些数据的拟合能力,从而更准确地进行寿命预测和可靠性分析。采用极大似然估计法对新提出的分布进行参数估计,并通过牛顿迭代法求解复杂的参数估计问题。对选定样本数据进行估计求解,并将模型与实际观测数据进行了对比。模型的拟合优度通过AIC和BIC值与其他模型进行了比较,结果表明新提出的扩展广义指数几何分布在拟合观测数据方面表现出较好的效果。Based on the generalized exponential geometric distribution,a new life distribution is proposed by extending the parameters.This extension is intended to enhance the model’s ability to fit some data,so that life prediction and reliability analysis can be performed more accurately.Parameter estimation for the newly proposed distribution is performed by using the method of maximum likelihood estimation,and the complex estimation issues are solved through Newton’s iterative method.Sample data are estimated and compared with actual observational data.The model’s goodness of fit is compared with that of other models through the AIC and BIC values,and the results indicate that the newly proposed extended generalized exponential geometric distribution performs well in fitting observational data.

关 键 词：指数几何分布 广义指数几何分布 极大似然估计 牛顿迭代法 

分 类 号：O21[理学—概率论与数理统计]