A Scenario Implementation in R for SubtypeDiscovery Examplified on Chemoinformatics Data

We developed a methodology that both facilitates and enhances the search for homogeneous subtypes in data. We applied this methodology to medical research on Osteoarthritis and Parkinson’s Disease and to chemoinformatics research on the chemical structure of molecule profiles. We release this method...

Full description

Saved in:
Bibliographic Details
Published inLeveraging Applications of Formal Methods, Verification and Validation Vol. No. 17; pp. 669 - 683
Main Author Steffen, Bernhard
Format Book Chapter
LanguageEnglish
Published The Netherlands Springer 2008
Springer Berlin Heidelberg
SeriesCommunications in Computer and Information Science
Subjects
3D graphics & modelling
Association Table
Bayesian Information Criterion
Bayesian Information Criterion Score
Cluster Membership
Cluster Result
Data structures
Systems analysis & design
Online AccessGet full text

Cover

More Information
Summary:We developed a methodology that both facilitates and enhances the search for homogeneous subtypes in data. We applied this methodology to medical research on Osteoarthritis and Parkinson’s Disease and to chemoinformatics research on the chemical structure of molecule profiles. We release this methodology as the RSubtypeDiscovery package to enable reproducibility of our analyses. In this paper, we present the package implementation and we illustrate its output on molecular data from chemoinformatics. Our methodology includes different techniques to process the data, a computational approach repeating data modelling to select for a number of subtypes or a type of model, and additional methods to characterize, compare and evaluate the top ranking models. Therefore, this methodology does not solely cluster data but it also produces a complete set of results to conduct a subtype discovery analysis.
ISBN:9783540884781
3540884785
ISSN:1865-0929
1865-0937
DOI:10.1007/978-3-540-88479-8_48