A novel training method has been proposed for increasing efficiency and generalization of support vector machine (SVM). The efficiency of SVM in classification is directly determi...
This work deals with the application of kernel methods to structured relational settings such as semantic knowledge bases expressed in Description Logics. Our method integrates a n...
Classifiers favoring sparse solutions, such as support vector machines, relevance vector machines, LASSO-regression based classifiers, etc., provide competitive methods for classi...
The problem of model selection for support vector machines (SVMs) is considered. We propose an evolutionary approach to determine multiple SVM hyperparameters: The covariance matr...
Background: Accurate identification of protein domain boundaries is useful for protein structure determination and prediction. However, predicting protein domain boundaries from a...