放射状角膜切开术预测系统中数学模型的建立

MATHEMATICAL MODEL FOR PREDICTING SYSTEM OF RADIAL KERATOTOMY

本文运用逐步回归方法建立放射状角膜切开术预测系统的非线性数学模型 ,分析各因素对手术效果的影响。该模型基于 10个参数 ,分别为 :年龄、球镜、角膜屈光 1(垂直方向屈光 )、角膜屈光 2 (水平方向屈光 )、角膜周边厚度、角膜直径 1(垂直方向直径 )、角膜直径 2 (水平方向直径 )、弹性眼差、视区、切口刀数 ,所要预测的值为角膜屈光改变值。应用友谊医院提供的眼科病人手术两年后复查的数据 ,提取参数齐全的数据 ,剔除可疑数据 ,得到 5 2 3组可用数据。用C语言编程对数据进行计算 ,分析各因素、各因素平方项及各因素间交互作用对角膜屈光改变值的影响 ,分别在显著性水平为 0 .2 5、0 .1、0 .0 5、0 .0 1下进行显著性检验 ,得到两种不同结果。两个模型预测值与实际手术值的误差小于 2D的百分比分别为 92 .6 5 %和 91.43%。

Stepwise regression analysis was used to build mathematical model of radial keratotomy(RK), for analyzing factors influencing on the result, and predicting the RK result. The model based on 10 factors including age, sphere, corneal level refraction, corneal vertical refraction, corneal depth(4 sites), eye axial length, corneal level diameter, pressure of eye, diameter of clear zone, and number of incisions. What we need to predict was the change, of corneal refraction. We based on the data of Friendshin Hospital Eyebank two years after the patients received RK. Using Chauvenet method abnormal data of 523 eyes were analyzed for building model. We analyzed data under four different signiticant, levels (0.25,0.1,0.05 and 0.01) and got two kinds of results. The difference between actual and predicted value within 2D was 92.65% or 91.43%.