pitprops14: Pitprops Correlation Data for 14 Variables

In whitening: Whitening and High-Dimensional Canonical Correlation Analysis

Pitprops Correlation Data for 14 Variables

Description

Pit prop timber is used in construction to build mining tunnels. The pitprops14 data is described in Jeffers (1967) and is a correlation matrix that was calculated from measuring 14 physical properties of 180 pit props made from wood from Corsican pines grown in East Anglia, UK.

Usage

data(pitprops14)

Format

A correlation matrix of dimension 14 times 14.

Details

The 14 variables are described in Jeffers (1967) as follows:

topdiam: the top diameter of the prop in inches; length: the length of the prop in inches; moist: the moisture content of the prop, expressed as a percentage of the dry weight; testsg: the specific gravity of the timber at the time of the test; ovensg: the oven-dry specific gravity of the timber; rinotop: the number of annual rings at the top of the prop; ringbut: the number of annual rings at the base of the prop; bowmax: the maximum bow in inches; bowdist: the distance of the point of maximum bow from the top of the prop in inches; whorls: the number of knot whorls; clear: the length of clear prop from the top of the prop in inches; knots: the average number of knots per whorl; diaknot: the average diameter of the knots in inches; maxcs: the maximum compressive strength in lb per square inch.

Source

The data set is printed in Jeffers (1967) in Table 2 and Table 5.

References

Jeffers, J. N. R. 1967. Two case studies in the application of principal component analysis. JRSS C (Applied Statistics) 16: 225-236. <DOI:10.2307/2985919>

Examples

# load whitening library
library("whitening")

# load pitprops14 data set
data(pitprops14)
colnames(pitprops14)

# correlation matrix for the first 13 variables
pitprops13 = pitprops14[1:13, 1:13]

# correlation loadings for PCA whitening
Psi = whiteningLoadings(pitprops13, "PCA")$Psi

# corresponding explained variation
Psi.explained = explainedVariation(Psi)

# the first six whitened variables account for 87% of the variation
cumsum(Psi.explained)/13*100

whitening documentation built on June 7, 2022, 5:10 p.m.