矩阵维度对Kappa作弊甄别方法性能的影响

The Influences of Matrix Dimension on the Performance of Kappa

本文通过对Kappa作弊甄别方法中顺序量表的44矩阵算法进行优化,研究矩阵维度的变化对Kappa方法甄别率及I型错误率的影响,并提出用于比较甄别性能的KX系数及基于矩阵维度选择算法的KappaX作弊甄别方法。

This article investigates the effectiveness, caused by the change of dimensions, to the detection rate and Type I error rate of Kappa approach. The approach used in this study is to optimize the 4 ＊ 4 matrix calculating method for ordinal scaling of Kappa. This article also presents KX index, a new index for comparing the detecting efficiency, and the detecting procedure for Kappa-X which chooses the calculating method based on the dimensions of the matrix.