摘要
该文对应用BP神经网络和降维法相结合在 0 - 1记分模式下估计项目参数和考生能力的方法作了概述 。
The item parameters and the person ability parameter in Logistic Models based on IRT are estimated with Back-Propagation Neural Network. The dimension of scoring matrices X is reduced by using scoring rate, passing rate, coefficient of correlation, guess rate when estimating corresponding ability parameter and item parameters. The results drawn from Monte Carlo study show that the item parameters estimation is more precise than the current prevailing softwares.The focus of this paper is to discuss how to apply this estimation method to practice.
出处
《江西师范大学学报（自然科学版）》
CAS
2004年第6期485-488,共4页
Journal of Jiangxi Normal University(Natural Science Edition)
基金
国家自然科学基金 (6 0 2 6 30 0 5 )
全国教育科学规划重点课题 (DBB0 1 0 5 0 1 )
江西省自然科学基金 (0 4 1 1 0 2 1 )资助项目
关键词
降维法
考生
IRT
实际
能力
记分
模式
参数估计
对应
Item Response Theory
item parameter estimation
Back-Propagation neural network
reduction of dimension
application
