標題: | 轉移機率含因子的分析方法 Analysis of Transition Probability with Covariates |
作者: | 艾雪芳 Hsueh-Fang Ai 彭南夫 Nan-Fu Peng 統計學研究所 |
關鍵字: | 退化性視網膜黃斑部病變;轉移機率;條件馬可夫鏈;ARM;transition probablity;conditional Markov Chain |
公開日期: | 2006 |
摘要: | 在過去的文獻[6]中提供了很多針對單一比率(發生率、惡化率…)的統計方法,而今我們想要發展的分析方法是針對含有風險因子的轉移資料(transition data with risk factors). 在條件馬可夫模型下我們可以解決資料不獨立的問題。
使用這個方法,我們需要足夠大的樣本所以我們把一些連續型的變數重新編碼,且用一些檢定找出比較有影響的因子,來建立統計模型。
在模型下,我們使用拔靴法(bootstrap method)來估計每個參數的信賴區間,據此來檢定各參數是否顯著。
最後我們可以看到這個疾病所有歷程上的轉移機率。所以我們可以做不同的比率(發生率、惡化率…)在不同的風險因子之間的比較。在本研究中,我們使用退化性視網膜黃斑部病變(ARM)的資料來展示這個方法,且提供之前對ARM 資料分析的回顧。 In Huang et al [6] the single rate research is provided by a lot of statistical methods. Now we want to develop a new method to analyze the transition data with risk factors. With the conditional Markov model we can solve the dependent data question. Use this method we need a large “enough” sample in each cell of transition probability. So we recode the continuous factors and use some tests to find more influential factors. Under the model, bootstrap method can help us to construct a confidence interval for each parameter. Thus we still can test the parameter if it is significant. The results include all the transition probability of the process of the disease. Then we can compare the rates among the factor. In this report we use ARM (age-related maculopathy) data to demonstrate the method and provide the previous research in analysis of ARM data. |
URI: | http://140.113.39.130/cdrfb3/record/nctu/#GT009426522 http://hdl.handle.net/11536/81462 |
Appears in Collections: | Thesis |
Files in This Item:
If it is a zip file, please download the file and unzip it, then open index.html in a browser to view the full text content.