Embedded cluster modelling - a novel method for analyzing embedded data sets
A novel method, called embedded cluster modelling, is proposed for the derivation of classification models for embedded data sets. These are data sets in which the chemicals are divided into two classes and in which one class of chemicals is found to cluster along one or more variables, forming an 'embedded cluster' surrounded by the 'diffuse cluster' of chemicals in the other class. The combined use of ECM and cluster significance analysis - a method which identifies variables along which clustering is statistically significant - is illustrated by their application to a data set of methacycline datasets.
Bibliographic Reference: Article: Quantitative Structure Activity Relationships, (1999), 229-235
Record Number: 199911413 / Last updated on: 1999-10-01
Original language: en
Available languages: en