odregress: Orthogonal Distance Regression.

In thartbm/handlocs: Download and analyse 2D reaching and hand localization data

Orthogonal Distance Regression.

Description

Perform (multiple) orthogonal distance regression.

Usage

odregress(x, y)

Arguments

x	A 'predictor' matrix (variables in columns, N observations in rows).
y	Column vector (N rows) with 'dependent' variable.

Details

This function is copied from the 'pracma' package by Hans W. Borchers: https://CRAN.R-project.org/package=pracma

For the 2D case, a pearson correlation coefficient can be returned. No usable coefficient of determination seems to exist for other cases.

Value

A list with: 'coeff' (coefficients), 'yfit' (fitted values), 'err' (errors), 'resd' (residuals), 'ssq' (sum of squares) and 'normal' (normal).

Examples

# we get data from the iris data set that is pretty correlated
# but since sepal length does not cause petal length,
# we use orthogonal distance regression:
x <- as.matrix(iris[101:150,1])
y <- as.matrix(iris[101:150,3])
odr <- odregress(x,y)

plot(x,y)
abline(a=odr$coeff[2], b=odr$coeff[1], col='red')

# compare to ordinary least squares:
ols <- lm(y ~ x)
abline(a=ols$coefficients[1], b=ols$coefficients[2], col='blue')

thartbm/handlocs documentation built on Feb. 18, 2025, 10:53 p.m.