calculate_z_scores: Calculate z scores for outliers

In skinnider/modern: Model-free outlier detection for robust network inference

Description

Convert a matrix of autocorrelations to a matrix of Z scores, such that the matrix has a mean of zero and a standard deviation of one. If the matrix contains missing values, this scaling is performed for each group of columns with equivalent numbers of missing values separately. Optionally, if there are many possible numbers of missing values, the z score can be calculated for approximately equal sized bins of missing value counts using the bins parameter.

Usage

calculate_z_scores(autocor, bins = NA)

Arguments

autocor	the matrix of autocorrelations
bins	optionally, the number of bins into which to group nodes on the basis of the number of observations

Value

a matrix with identical dimensions to the input matrix, containing the autocorrelation Z score assigned to each non-missing observation

skinnider/modern documentation built on Feb. 20, 2020, 1:52 p.m.