文章基本信息

标题：Moment-Based Density Approximation Techniques as Applied to Heavy-tailed Distributions
本地全文：下载
作者：John Sang Jin Kang ; Serge Provost ; Jiandong Ren 等
期刊名称：International Journal of Statistics and Probability
印刷版ISSN：1927-7032
电子版ISSN：1927-7040
出版年度：2019
卷号：8
期号：3
页码：1-23
DOI：10.5539/ijsp.v8n3p1
出版社：Canadian Center of Science and Education
摘要：Purpose To investigate factors affecting persistent diplopia after surgical repair of isolated inferior orbital wall fractures. Methods Thirty-three patients who underwent surgical repair of isolated inferior orbital wall fractures in Inha University Hospital Ophthalmology Department from 2014 to 2017 were enrolled in this study. The authors examined facial computed tomography, diplopia, extraocular muscle movement, and Hertel's exophthalmometer before and 6 months after surgery. The diplopia which was not recovered even at 6 months postoperatively was defined as persistent diplopia. Multivariable logistic regression analyses were performed on parameters that were found to be related to persistent diplopia using univariable logistic regression analyses. Results Univariable regression analysis showed that preoperative ocular motility limitation, preoperative diplopia, the type of fracture, the number of contacts with the fracture site and extraocular muscle (EOM), and EOM tenting were associated with persistent postoperative diplopia. Multivariable regression analysis using the previously mentioned five parameters showed 28.3-fold and 17.4-fold greater probabilities of diplopia after surgery in preoperative diplopia and EOM tenting, respectively ( p = 0.023). Conclusions Preoperative diplopia and EOM tenting were associated with persistent postoperative diplopia. These parameters were predictors of persistent diplopia in eyes with isolated inferior orbital wall fractures.
关键词：density approximation; heavy-tailed distributions; exponential tilting; integrated squared error; symmetrization technique; risk measures