小样本标准参照测验中项目参数估计的GRNN方法  

Item Parameter-estimation Method of GRNN in Criterion-referenced Test with Small Samples

作　　者：余嘉元[1] 汪存友[1] 

机构地区：[1]南京师范大学教育科学学院,江苏南京210097

出　　处：《广西师范大学学报（自然科学版）》2006年第4期107-110,共4页Journal of Guangxi Normal University:Natural Science Edition

基　　金：全国教育科学"十五"规划项目教育部重点课题(DBB010511);江苏省教育厅自然科学研究项目(02KjB180005)

摘　　要：针对标准参照测验题库建设中的项目参数估计方法所存在的问题,分析了先前的GRNN方法估计项目参数研究中所存在的不足,对原有方法进行了改进,并以二值记分下的单参数Logistic模型为研究对象进行了模拟实验。结果表明,在考生样本很小的情况下,当训练项目达到一定数目时,将会得到比较实用的GRNN估计结果。方法具有较高的实用价值。To solve the problems of some item parameter-estimation when constructing item bank for criterion-referenced test ,as well as the drawback exposure of earlier research on how to apply GRNN to estimate item parameter within item response theory, the earlier method has been improved here. Simulation experiments were carried out, based on the item response model of one parameter Logistic model conditioned on binomial scoring. And the result shows that the with small number of examinee,practical estimation error can be obtained especially when the training item samples increase to a certain number, which indicates that the method proposed here is valuable.

关 键 词：标准参照测验 广义回归神经网络 项目反应理论 参数估计 计算机模拟 小样本 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] TP391.7[自动化与计算机技术—控制科学与工程]