The results of the 2006 ECML/PKDD Discovery Challenge suggest that semi-supervised learning methods work well for spam filtering when the source of available labeled examples diff...
We propose a novel privacy-preserving support vector machine (SVM) classifier for a data matrix A whose input feature columns are divided into groups belonging to different entiti...
—The paper discusses the end-to-end QoS provisioning for scalable video streaming traffic delivery over heterogeneous IP/UMTS networks. A prototype architecture is proposed, and ...
Thomas Pliakas, George Kormentzas, Charalabos Skia...
Abstract— We present and evaluate a novel constraintbased model transformation approach that implements a preservation-centric view. The proposed framework comprises formal prese...
— Many future ubiquitous devices will be able to determine their precise physical position using different types of localization techniques (e.g. GPS antennas for outdoor-, RFID ...