查詢結果分析
來源資料
頁籤選單縮合
| 題 名 | Random-Effects Mixture Model with AR(1) Dependence for Classifying Longitudinal Clinical Data=具AR(1)相關之混合隨機效應模型用於臨床試驗數據分類 |
|---|---|
| 作 者 | 林資荃; 蔡貴鳳; | 書刊名 | 中國統計學報 |
| 卷 期 | 56:1 2018.03[民107.03] |
| 頁 次 | 頁7-24 |
| 分類號 | 319.5 |
| 關鍵詞 | 長期資料; 混合隨機效應模型; AR(1)過程; ECME演算法; Longitudinal data; Random-effects mixture model; AR(1) process; ECME algorithm; |
| 語 文 | 英文(English) |
| 中文摘要 | 長期資料分析中,常見的統計分析方法為混合固定和隨機效應的線性模式。此模式中通常假設隨機效應服從單一常態分配,即受試者皆來自同質性母體。然而,單一常態分配假設可能過度簡化受試者間變異性,於是考量隨機效應服從混合常態分配可被擴展用來處理受試者間異質性,由於數學和計算易處理性,這些方法仍假設受試者內部誤差為獨立。本文章將受試者內部的誤差擴展為具有AR(1)相關性,提出一個蒙地卡羅ECME演算法用來估計具AR(1)相關之混合隨機效應模型之參數值。最後,使用模擬資料與實際例子來闡述所提出的方法。 |
| 英文摘要 | Linear mixed-effects (LME) models have been widely used for longitudinal data analysis as it can account for both fixed and random effects, while simultaneously incorporating the variation on both within and between subjects. The random effects of LME models are commonly assumed to follow a single normal distribution, which may oversimplify the between-subject variability. Mixtures of linear mixed-effects (MLME) models have been proposed to accommodate the heterogeneity problem of random effects. However, the within-subject errors in MLEM models are assumed to be independent for mathematical and computational tractability. In this article, we consider an AR (1) process to address the possible serial autocorrelation among within-subject errors in MLME models. We describe a flexible hierarchical representation of the considered model and develop an efficient Expectation-Conditional Maximization Either (ECME) algorithm for carrying out maximum likelihood estimation. The performance of the proposed method is illustrated on both real and simulated datasets. |
本系統中英文摘要資訊取自各篇刊載內容。