本研究計畫提出一個依據反?迴歸的無母?方法，用以分析高維?長期追蹤資?。在維?縮減以外，我們研究變?選取。透過維?縮減技巧，一個資?調適型的搜尋方法被提出用以模型配適。模擬研究與實際資?結果將用以?明本法表現。 In this project, we propose a nonparametric method based on sliced inverse regression for analyzing high-dimensional longitudinal data. Variable selection along with dimension reduction is studied. A data-adaptive searching method via our sufficient dimension reduction technique for model fitting is presented. Several simulation and empirical results are reported for illustration.
