Premium
Selection of Characteristics for Numerical Classification of Soils
Author(s) -
Sarkar P. K.,
Bidwell O. W.,
Marcus L. F.
Publication year - 1966
Publication title -
soil science society of america journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.836
H-Index - 168
eISSN - 1435-0661
pISSN - 0361-5995
DOI - 10.2136/sssaj1966.03615995003000020035x
Subject(s) - weighting , soil water , selection (genetic algorithm) , correlation coefficient , moment (physics) , mathematics , product (mathematics) , correlation , soil classification , statistics , soil science , computer science , environmental science , artificial intelligence , geometry , physics , classical mechanics , acoustics
Eighteen hundred and thirty product‐moment correlation coefficients were used to select characteristics from 61 given soil characteristics for numerical classification by electronic computer of 26 soils representing nine Orders of the new soil classification system (7th Approximation). Correlated characters were eliminated successively to 51, 40, 33, and 22 soil characteristics according to decreasing levels of the product‐moment correlation coefficient over soils. The correlation method is an objective procedure for eliminating members of groups of similar characteristics that might produce undue weighting in the numerical classification. It is superior to the heretofore used, and often criticized, method of selection based on subjective evaluation of the characteristics in soil classification.