【摘要】：Meta-分析方法是一種對(duì)具有相同研究目的的多個(gè)研究結(jié)果進(jìn)行合并,并綜合評(píng)價(jià)結(jié)果的統(tǒng)計(jì)學(xué)方法,該方法在循證醫(yī)學(xué)領(lǐng)域有著廣泛的應(yīng)用.當(dāng)某事件概率發(fā)生很小時(shí),Meta-分析經(jīng)常會(huì)遇到稀疏數(shù)據(jù)問(wèn)題,針對(duì)稀疏數(shù)據(jù)的處理目前主要有兩種方法:一是對(duì)其進(jìn)行連續(xù)性修正;二是利用貝葉斯方法進(jìn)行分析.本文主要是基于貝葉斯方法進(jìn)行Meta-分析,給出了 Jeffreys無(wú)信息先驗(yàn)Beta(1/2,1/2)時(shí)的鏈接分布函數(shù):并將其推廣,給出了先驗(yàn)分布為一般貝塔分布Beta(α,β)時(shí)的鏈接分布函數(shù):最后我們給出了數(shù)值模擬結(jié)果,并討論了在可選擇的范圍內(nèi)t值的變化對(duì)試驗(yàn)效果的影響.
[Abstract]:Meta- analysis is a statistical method that combines and synthetically evaluates the results of multiple studies with the same purpose. This method is widely used in the field of evidence-based medicine. The problem of sparse data is often encountered when the probability of an event is very small. There are two main methods to deal with sparse data: one is to modify it continuously; the other is to use Bayesian method to analyze it. In this paper, the Meta- analysis based on Bayesian method is used, and the link distribution function of Jeffreys Beta (1 / 2 / 1 / 2) is given. The link distribution function when the prior distribution is a general Beta distribution Beta (偽, 尾) is given. Finally, the numerical simulation results are given, and the influence of the change of t value on the test results is discussed.
【學(xué)位授予單位】：山東師范大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：C81

【參考文獻(xiàn)】

相關(guān)期刊論文 前3條

1 曾憲濤;冷衛(wèi)東;李勝;郭毅;王樸;;如何正確理解及使用GRADE系統(tǒng)[J];中國(guó)循證醫(yī)學(xué)雜志;2011年09期

2 王丹;翟俊霞;牟振云;宗紅俠;趙曉東;王學(xué)義;顧平;;Meta分析中的異質(zhì)性及其處理方法[J];中國(guó)循證醫(yī)學(xué)雜志;2009年10期

3 王珍;張紅;潘云;;Meta-分析固定效應(yīng)模型及其擴(kuò)展應(yīng)用[J];中國(guó)組織工程研究與臨床康復(fù);2008年39期

，

本文編號(hào)：2220351