一種基于貝葉斯的廣義 Pareto 分布變點估計

打開文本圖片集

摘要：文章研究廣義 Pareto 分布單變點的估計問題，利用貝葉斯方法對廣義 Pareto分布變點進行估計，模擬結(jié)果表明，貝葉斯方法能獲得更好的效果。同時將貝葉斯方法與基于 KL 散度似然比統(tǒng)計量的極大似然法比較，得出貝葉斯方法效果更好。

關(guān)鍵詞：廣義 Pareto 分布;貝葉斯估計;極大似然估計

中圖分類號：0212.8   文獻標識碼：A   文章編號：1009-3583（2024）-0108-05

A Bayesian Based Generalized Pareto Distribution Change Point Estimation

WANG Guo-qin1 ， WU You-fu2* ， XU Ting1 ， OU Yong-ling1

（1.School of Data Science and Information Engineering， Guizhou Minzu University;2. Guizhou Communication Vocational College， Guiyang 550025， China;）

Abstract： The article studies the estimation problem of the single change point of the generalized Pareto distribution， and uses Bayesian methods to estimate the change point of the generalized Pareto distribution. Simulation results show that Bayesian methods can achieve better results. At the same time， comparing the Bayesian method with the maximum likelihood method based on KL divergence likeli- hood ratio statistic， it is found that the Bayesian method has better performance.

keywords： generalized Pareto distribution; Bayesian estimation; maximum likelihood estimation

在統(tǒng)計學中一個比較熱門的研究問題就是變點問題，它被用于經(jīng)濟學、氣象學、醫(yī)學等領(lǐng)域。（剩余4786字）