【摘要】：伴隨著科技和醫(yī)療衛(wèi)生水平的不斷提高,醫(yī)學統(tǒng)計等臨床試驗研究發(fā)現(xiàn),對于那些反復(fù)發(fā)生,并且在人們眼中似乎是不可能被治愈的疾病,近年來發(fā)現(xiàn)有疑似治愈個體的存在。在復(fù)發(fā)事件過程中,我們將考慮治愈個體的影響,推廣混合治愈模型,使之能夠更為準確合理地解釋復(fù)發(fā)事件過程。在第二章中,本文在復(fù)發(fā)事件數(shù)據(jù),半?yún)?shù)比率模型基礎(chǔ)之上,利用Logistic模型回歸治愈率部分,提出一類含有有治愈個體的混合治愈模型,來刻畫協(xié)變量對事件復(fù)發(fā)率的影響。同時闡述模型中未知參數(shù)的一種估計方法,并詳細證明這些估計的大樣本性質(zhì)。接著利用數(shù)值模擬的方法,量化地說明這些估計在有限樣本下是行之有效的,最后讓我們的模型及方法,運用到一組真實的膀胱癌數(shù)據(jù)中。在第三章中,同樣是在復(fù)發(fā)事件過程中,我們建立一類更為一般的混合治愈模型,該模型包含了一大類混合治愈回歸模型,第二章中混合治愈模型是其特殊情形。與估計方程的思想一致,我們敘述模型參數(shù)的估計方法,并詳細地證明了這些估計的大樣本性質(zhì)。最后,我們將總結(jié)這幾項研究,并提出一些后續(xù)的研究計劃。
[Abstract]:With the continuous improvement of science and technology and medical and health standards, medical statistics and other clinical trials have found that for those diseases that occur repeatedly and seem to be impossible to cure in people's eyes, the existence of suspected cured individuals has been found in recent years. In the process of recurrent events, we will consider the influence of cured individuals and extend the mixed cure model to explain the recurrence events more accurately and reasonably. In the second chapter, on the basis of recurrence event data and semi-parametric ratio model, using Logistic model regression cure rate part, a kind of mixed cure model with cured individuals is proposed to describe the influence of covariable on event recurrence rate. At the same time, an estimation method of unknown parameters in the model is presented, and the large sample properties of these estimators are proved in detail. Then the numerical simulation method is used to quantitatively show that these estimates are effective in a limited sample. Finally, let our model and method be applied to a set of real bladder cancer data. In the third chapter, we also establish a more general mixed cure model, which includes a large class of mixed cure regression models. In chapter 2, the mixed cure model is a special case. In accordance with the idea of the estimation equations, we describe the estimation methods of the model parameters, and prove in detail the large sample properties of these estimates. Finally, we will summarize these studies and put forward some further research plans.
【學位授予單位】：海南師范大學
【學位級別】：碩士
【學位授予年份】：2015
【分類號】：O212.1

【相似文獻】

相關(guān)碩士學位論文 前1條

1 曾小鳳;復(fù)發(fā)事件過程中混合治愈模型及其統(tǒng)計分析[D];海南師范大學;2015年

，

本文編號：2364403