In this paper, we propose a technique for segmenting visual textures using features extracted from the reponses of Ga,bor filters, appropria.tely selectecl to be tuned to texture ...
In many pattern recognition applications, high-dimensional feature vectors impose a high computational cost as well as the risk of "overfitting". Feature Selection addre...
Abstract. Computational analysis of mass spectrometric (MS) proteomic data from sera is of potential relevance for diagnosis, prognosis, choice of therapy, and study of disease act...
Elena Marchiori, Connie R. Jimenez, Mikkel West-Ni...
In this paper, we compare case-based spam filters, focusing on their resilience to concept drift. In particular, we evaluate how to track concept drift using a case-based spam fi...
Feature selection is one of key issues related with data pre-processing of classification task in a data mining process. Although many efforts have been done to improve typical fea...