基于FAST-MCD算法的异常成绩检测研究

Research on the Abnormal Grade Detection Based on the FAST-MCD Algorithm

在教学过程中学生异常成绩的检测,有助于及时地发现学生学习中的问题,针对性的调整教学方法,从而改进教学。本文假定学生的每一次测验成绩符合高斯分布,利用最小协方差估计(MCD)检测学生成绩中的异常数据,通过采用FAST-MCD算法提升计算效率。实验发现:依据经验给出异常成绩比例,基于FAST-MCD算法的异常成绩检测能更快、更具有鲁棒性地检测异常数据。

In the process of teaching,the detection of students’abnormal grade is helpful to discover students’learning problems in time,and adjust teaching methods accordingly,so as to improve teaching.In this paper,we assume that the students'grades in each test fitting to Gaussian distribution,use the minimum covariance estimation(MCD)to detect the abnormal data in the students’grades,and use the FAST-MCD algorithm to improve the computational efficiency.The experimental results show that:according to the experi⁃ence,the abnormal grade detection based on FAST-MCD algorithm can detect faster and more robustly.