This correlation matrix was published in Jeffers (1967) and was calculated from 180 observations. The 13 variables were used as explanatory variables in a regression problem which arised from a study on the strength of pitprops cut from home-grown timber.
data(pitpropC)Its a correlation matrix of 13 variables which have the following meaning:
| [,1] | TOPDIAM | Top diameter of the prop in inches |
| [,2] | LENGTH | Length of the prop in inches |
| [,3] | MOIST | Moisture content of the prop, expressed as a percentage of the dry weight |
| [,4] | TESTSG | Specific gravity of the timber at the time of the test |
| [,5] | OVENSG | Oven-dry specific gravity of the timber |
| [,6] | RINGTOP | Number of annual rings at the top of the prop |
| [,7] | RINGBUT | Number of annual rings at the base of the prop |
| [,8] | BOWMAX | Maximum bow in inches |
| [,9] | BOWDIST | Distance of the point of maximum bow from the top of the prop in inches |
| [,10] | WHORLS | Number of knot whorls |
| [,11] | CLEAR | Length of clear prop from the top of the prop in inches |
| [,12] | KNOTS | Average number of knots per whorl |
| [,13] | DIAKNOT | Average diameter of the knots in inches |
Jeffers (1967) replaced these 13 variables by their first six principal components. As noted by Vines (2000), this is an example where simple structure has proven difficult to detect in the past.
Jeffers, J.N.R. (1967) Two case studies in the application of principal components analysis. Appl. Statist. 16, 225--236.
Vines, S.K. (2000) Simple principal components. Appl. Statist. 49, 441--451.
data(pitpropC)
symnum(pitpropC)
Run the code above in your browser using DataLab