Application of discriminant analysis in design review
- 1 February 1973
- journal article
- Published by American Geophysical Union (AGU) in Water Resources Research
- Vol. 9 (1) , 93-102
- https://doi.org/10.1029/wr009i001p00093
Abstract
Discriminant analysis is a multivariate technique that uses a group of variables (discriminators) to obtain maximum separation between different groups of data. The discriminant function, a linear combination of the discriminators, obtained by this analysis may be used to assign an individual observation to the group that it most nearly resembles. Discriminant analysis was used to analyze factors related to the adequacy of standard slope protection on small dams. The structures were divided into two groups: one group that visibly needed additional slope protection and one group that appeared to be designed satisfactorily. The analysis showed that a linear combination of the fetch length, the surface area to length ratio (average width), and the plasticity index of the surface material on the dam could be used to distinguish between the two groups about 85% of the time. The equation defined by these parameters was scaled for cost and number of structures, and can be used to assign structures to the two groups such that the additional protection can be most economically provided.This publication has 3 references indexed in Scilit:
- When is it safe to extend a prediction equation?—An answer based upon factor and discriminant function analysisWater Resources Research, 1967
- Statistical Prediction by Discriminant AnalysisPublished by Springer Nature ,1962
- Statistical Prediction by Discriminant AnalysisPublished by Springer Nature ,1962