Guest editors' introduction: Special issue of selected papers from ECML PKDD 2009

A brief overview of the 14 papers selected for the special issue of Data Mining and Knowledge Discovery (PKDD) and the European Conference on Machine Learning (ECML), is presented. Joachims and Yu in their paper, 'Sparse kernel SVMs via cutting-plane training' advance the state of the art in efficient induction and evaluation of SVM models. Roth and Samdani present a paper titled, 'Learning multi-linear representations of distributions for efficient inference', presenting a method of speeding up probabilistic inference by representing discrete probability distributions using multilinear forms. Liu and coresearchers present 'A self-training approach to cost sensitive uncertainty sampling', demonstrating how traditional forms of uncertainty sampling in active learning fail in asymmetric misclassification cost scenarios and propose an remedy technique based on self-training. Zhao and coworkers present 'A fast ensemble pruning algorithm based on pattern mining process'.