兩面迴歸其切割面的貝氏估計

Abstract

本論文的目標在估計兩面多維線性迴歸的切割面.針對三種不同的事前假 設 ,我們採取貝氏的方法去估計切割面.最後,以蒙地卡羅的模擬實驗來說 明利用貝氏方法所得結果是個很好的估計方法. We consider the estimation problem of change hyperplane for a two phase multiple linear regression model. A Bayesian technique has been proposed for three underlying assumption sets. Also, a Monte Caro study show that the proposed Bayesian technique has very good performance under two criterions.