Croatica Chemica Acta, Vol. 62 No. 3, 1989.
Original scientific paper
Classification of Steroid Partition Coefficients by a Pattern Recognition Technique
A. J. Harget
; Dept. of Computer Science and Applied Mathematics, Aston University, Aston Triangle, Birmingham B4 7ET, UK
D. S. Ellis
; Dept. of Computer Science and Applied Mathematics, Aston University, Aston Triangle, Birmingham B4 7ET, UK
Abstract
A pattern recognition technique, the linear learning machine
method, has been used to determine structure-activity relationships
for certain steroids. The steroids used in this study were classified into two categories according to their observed partition coefficient and a correlation made with certain substructural descriptors. The linear learning machine method was employed to calculate a suitable decision surface that would classify each steroid into its correct category. The resulting structure-activity relationship and the relative contributions of the various structural variables are discussed, and a comparison made with resultsobtained from a study using a different approach.
Keywords
Hrčak ID:
175803
URI
Publication date:
10.10.1989.
Visits: 916 *