Computers in chemistry V 0380 Representing Clusters Using a Maximum Common Edge Substructure Algorithm Applied to Reduced Graphs and Molecular Graphs. -(GARDINER*, E. J.; GILLET, V. J.; WILLETT, P.; COSGROVE DAVID A.; J. Chem. Inf. Model. (J. Chem. Inf. Comput. Sci.) 47 (2007) 2, 354-366; Dep. Inf. Stud., Univ. Sheffield, Sheffield S10 2TN, UK; Eng.) -Lindner 24-188

No abstract

In this study, a support vector machine (SVM) classifies real world data of cytogenetic signals measured from fluorescence in-situ hybridi:ation (FISH) images in order to diagnose genetic syndromes. The study implements the SVM structural risk minimization concept in searching for the optimal setting of the classifier kernel and parameters. We propose thresholding the distance of resied pafterns from ihe SVM separating hyperplane as a way of rejecting a percentage of the miss-classified patterns thereby alloiving reduction of the expected risk. Results show accurate performance of the SVM in classt$ing FISH signals in comparison to other state-of the-arr machine learning classijers, indicating the potential of an SVM-based genetic diagnosis systeni.

We extend the standard two-stage construction of the double-sided Golay-Rudin-Shapiro sequence to higher dimensions. We explicitly present the two-dimensional structure which is a convenient paradigm for all natural dimensions. We show also the three-dimensional four-symbol first-stage structure which we call proto-GRS and the final two-symbol Golay-Rudin-Shapiro structure. They may serve as models of disordered equicomposition alloys with some short range order of four and two components, respectively. Finally we show the essential features of the three- and four-dimensional structures.

In this study, a support vector machine (SVM)