cortrafo: cortrafo

In networktree: Recursive Partitioning of Network Models

Description

An influence function that transforms the response variables (y1, y2, y3...) into scores relevant to the correlations, means, and/or variances.

For example, in the case of correlations the variables are transformed into a matrix of (n^2-n)/2 columns (e.g., the number of total correlations), and i rows, where i is the # of observations of y1, where the mean of each vector is equal to the correlation between y1 and y2, y1 and y3, etc.

Used internally in when method="ctree".

Usage

cortrafo(data, weights, control, model, ...)

Arguments

data	a matrix or data
weights	not currently used
control	not currently used
model	can be any combination of c("correlation", "mean", "variance"). Scores are determined based on the specified characteristics
...	not currently used

networktree documentation built on Feb. 4, 2021, 5:06 p.m.