关于学生复杂连锁反应时间分布模型参数估计的一种方法

A Method of Parameter Estimation of Students'Response Time Distribution Model in the Complex Chain

介绍了在计算机局域网环境条件下进行形成性检测,以及收集学生群体的反应数据并加以处理的一般流程,导出关于总体参数t0、m和τ矩估计的非线性方程组.鉴于矩估计编程算法复杂,且对于超越方程组可能得不到满足精度的数值解,给出一种通过线性回归进行参数估计的方法,并对可能存在的异方差问题,给出使用加权最小二乘法估计总体参数的解决方案.

The paper presents the general process that collects and deals with the data of the students＇response time in the formative tests conducted under the condition of computer local area network,and it derives the group of nonlinear equations about the moment estimation of population parameters t0,mandτ.Since the moment estimation algorithm is very complex and the transcendental equations may not obtain the numerical solution that meets the accuracy,this paper discusses the method of parameter estimation through the linear regression.Aiming at the possible heteroscedastic problems,the paper gives the corresponsive solutions that can estimate population parameters by using the weighted least square method.