具有線性共變異結構的成長曲線模式

Abstract

這篇論文的主要目的是分析一些不等長（uubalanced）的成長曲線資料包括了生物的成長和科技的替代。我們從貝氏的觀點來看這問題，並使用Box–Cox轉換於不等長的成長曲線模式其變異矩陣為線性共變異結構。我們同時研究了參數的估計和未來值的預測。我們作了一些模擬結果並舉了實際資料來做分析。

The main burden of this article is to analyze unbalanced growth curve data including biological growth and technology substitutions. We treat such problems from a Bayesian point of view using an unbalanced growth curve model when the covariance matrix is Σ＝Σqz=θiGi, where θi is an unknown parameter, Gi is a known matrix, and the domain of θi satisfies Σqi=1θiGi being useful in statistics, while applying the Box-Cox power transformations on the observations. We study both estimation of parameters and prediction of future values. Some results are illustrated with several sets of real and simulated data.