CONFIDENCE INTERVALS ON LOW-ORDER POLYNOMIAL TREND SURFACES,
Abstract
Evaluation of trend srfaces fitted to map data in geology is largely substantive, supplemented by informal statistical test. Substantive elements include the degree to which the trend agrees with the underlying geological model; the statistical tests are based mainly on examination of the sums of squares associated with the fitted trend and with the deviations from the trend. The paper describes a formal method for generating upper and lower confidence surfaces about the fitted trend surface, as a basis for assessing the trend substantively, and for using the fitted surface as a predicting devide. The method is based on the covariance matrix of the polynomial coefficients; and example involving seemingly noisy map data is used for illustration and for comparison with less formal statistical analysis. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 12, 1963
- Accession Number
- AD0430317
Entities
People
- W. C. Krumbein
Organizations
- Northwestern University