頁籤選單縮合
題名 | 雙截斷資料下兩組存活函數檢定方法之探討=Testing the Equality of Two Survival Functions with Doubly Truncated Data |
---|---|
作 者 | 嵇允嬋; 蔡偉彥; 胡家玲; | 書刊名 | 中國統計學報 |
卷期 | 42:3 2004.09[民93.09] |
頁次 | 頁223-244 |
分類號 | 319.5 |
關鍵詞 | 雙截斷資料; 存活函數; 自我一致估計量; Doubly truncated data; Self-consistent estimator; Survival function; |
語文 | 中文(Chinese) |
中文摘要 | 關於雙截斷資料 (doubly truncated data) 存活函數估計之方法, 已有不少源自概似函數而推導出的演算法,如Thrnbull (1976) 利 用EM演算法所得的自我一致 (self-consistent) 估計式,Efron與 Petrosian (1999) 中列出由概似函數直接推導出的自我一致估計式,和他們提出以涉險個數為依據的演算法,這些方法所得的存活函數最大概似估計量皆是相同的。然而,目前尚未有學者對雙截斷資料提供兩組存活函數之無母數檢定方法。因此,於本文中將推廣Petroni與Wolfe (1994) 的想法,建議以雙截斷資料下兩組估計的存活函數之差為檢定統計量,並以概似函數推導出該檢定統計量的變異數。模擬驗證了所建議的檢定統計量之漸近抽樣分布皆為標準常態分布。除此之外,摸擬結果顯示,採用較多訊息所得的檢定統計量的檢定力較高。 |
英文摘要 | Recently, several algorithms, such as Thrnbull (1976) and Efron and Petrosian (1999), have been derived from the likelihood function to estimate the survival function based on doubly truncated data. However, the nonparametric methods for comparing two survival functions have not been developed. Therefore, this paper proposes two test statistics based on the differences between two estimated survival functions to test the equality of two survival functions for doubly truncated data. The comparative results from a simulation study are presented and the implementation of these methods to a real data set is presented. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。