Identify outliers in the multivariate distribution
在多元分布确定离群
译者:生物统计家园网 机器人LoveR
描述----------Description----------
A PCA model is fitted to data and two statistics as measures of extremity are calculated. These are the Hotelling t-square and DMODX, the first is a measure of how far away from the centre of the projection subspace the projection of the observation is. The second one measures how remote from the projection the actual observation is. SVD is done directly on the data matrix. The number of significant dimensions is defined as the number of eigenvalues greater than 1. Typically arrays are in different columns.
一个装有PCA模型数据和两下肢措施计算统计。这是霍特林丁字尺和DMODX的,首先是衡量投影的观察是从投影子空间的中心距离多远。第二个措施,如何远离投影的实际观察。 SVD的直接完成对数据矩阵。被定义为特征值大于1的若干重要方面。通常情况下,数组是在不同的列。
用法----------Usage----------
outlier(M)
参数----------Arguments----------
参数:M
matrix
矩阵
值----------Value----------
Dataframe with columns Hotelling and DMODX
dataframe与列霍特林和DMODX的