Premium
Generating hypotheses about molecular structure–activity relationships (SARs) by solving an optimization problem
Author(s) -
Ma Junshui,
Tong Christopher,
Liaw Andy,
Sheridan Robert,
Szumiloski John,
Svetnik Vladimir
Publication year - 2009
Publication title -
statistical analysis and data mining: the asa data science journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.381
H-Index - 33
eISSN - 1932-1872
pISSN - 1932-1864
DOI - 10.1002/sam.10040
Subject(s) - substructure , computer science , set (abstract data type) , context (archaeology) , data mining , optimization problem , data set , algorithm , artificial intelligence , engineering , paleontology , structural engineering , biology , programming language
This paper proposes a new automatic hypothesis‐generation algorithm for structure–activity relationship (SAR) rules, which is capable of investigating chemical compound activities in the context of multiple substructure interactions. The algorithm is formulated as an optimization problem based on a carefully selected criterion, APostDiff(s), and the globally optimal solution to the optimization problem can be obtained with a fast search algorithm developed using the data‐mining concept known as frequent set . Three public datasets are used to demonstrate the proposed method. Copyright © 2009 Wiley Periodicals, Inc. Analysis and Data Mining 1: 161‐174, 2009