Most of supervised learning algorithms assume the stability of the target concept over time. Nevertheless in many real-user modeling systems, where the data is collected over an ex...
Classifiers favoring sparse solutions, such as support vector machines, relevance vector machines, LASSO-regression based classifiers, etc., provide competitive methods for classi...
Abstract--Statistical pattern recognition techniques have recently been shown to provide a finer balance between misdetections and false alarms than the more conventional intrusion...
Background: Proteins are comprised of one or several building blocks, known as domains. Such domains can be classified into families according to their evolutionary origin. Wherea...
Elon Portugaly, Amir Harel, Nathan Linial, Michal ...
In this paper we embed evolutionary computation into statistical learning theory. First, we outline the connection between large margin optimization and statistical learning and s...