National Research Council of Canada. NRC Institute for Information Technology
Learning, Networks, and Statistics: Proceedings of the 1996 Workshop of the International School for the Synthesis of Expert Knowledge (ISSEK '96)
This study is concerned with whether it is possible to detect what information contained in the training data and background knowledge is relevant for solving the learning problem, and whether irrelevant information can be eliminated in pre-processing before starting the learning process. A case study of data pre-processing for a hybrid genetic algorithm shows that the elimination of irrelevant features can substantially improve the efficiency of learning. In addition, cost-sensitive feature elimination can be effective for reducing costs of induced hypotheses.