碩士 / 輔仁大學 / 統計資訊學系應用統計碩士班 / 104 / Maternal serum screening is the safest and common laboratory technique used for preliminary prenatal diagnosis of Down’s syndrome. Among the established maternal serum screening methods, the likelihood ratio (LR) model proposed by Cuckle et al. (1987) is the most commonly used screening method. However, many studies pointed out the variables of the age and maternal serum are not independent. Additionally, Zhong et al. (2011) found the logistic regression (LRG) method was better than LR method. In this thesis, we propose a nonlinear adjustment method based on LRG (NALRG) for Down’s syndrome screening. It is found that the LRG is a special member of NALRG. The best predictive NALRG model is determined by maximizing the efficiency ratio (ER) proposed by Baran et al. (2013). In our simulation, 1000 replicates were generated, and five-fold cross-validation was used to evaluate the performances of three different approaches. To determine the threshold, the methods proposed by Youden (1950) and Hosmer & Lemeshow (2004) were applied. The results showed that the NALRG is preferable to all its competitors in most cases. In addition, the results of cross-validation under nine simulation scenarios revealed the proposed NALRG outperforms all the other alternatives in terms of efficiency ratio.
| Identifer | oai:union.ndltd.org:TW/104FJU00506011 |
| Date | January 2016 |
| Creators | Yu,Po-Wei, 余柏暐 |
| Contributors | John Jen Tai, Chia-Ding Hou, 戴政, 侯家鼎 |
| Source Sets | National Digital Library of Theses and Dissertations in Taiwan |
| Language | zh-TW |
| Detected Language | English |
| Type | 學位論文 ; thesis |
| Format | 51 |
Page generated in 0.0022 seconds