Feature Construction and δ-Free Sets in 0/1 Samples
Given the recent breakthrough in constraint-based mining of local patterns, we decided to investigate its impact on feature construction for classification tasks. We discuss preliminary results concerning the use of the so-called δ-free sets. Our guess is that their minimality might help to collect important features. Once these sets are computed, we propose to select the essential ones w.r.t. class separation and generalization as new features. Our experiments have given encouraging results.
KeywordsFrequent Itemset Class Separation Feature Construction Interestingness Measure Viral Meningitis
Unable to display preview. Download preview PDF.
- 2.Boulicaut, J.-F., Crémilleux, B.: Simplest rules characterizing classes generated by delta-free sets. In: 22nd SGAI International Conference on Knowledge Based Systems and Applied Artificial Intelligence, ES 2002, pp. 33–46 (2002)Google Scholar
- 3.Becquet, C., Blachon, S., Jeudy, B., Boulicaut, J.F., Gandrillon, O.: Strong association rule mining for large gene expression data analysis: A case study on human SAGE data. Genome Biology 12 (2002)Google Scholar
- 4.Li, J., Li, H., Wong, L., Pei, J., Dong, G.: Minimum description length principle: Generators are preferable to closed patterns. In: Proceedings 21st National Conference on Artificial Intelligence. The AAAI Press, Menlo Park (2006)Google Scholar
- 5.Newman, D., Hettich, S., Blake, C., Merz, C.: UCI repository of machine learning databases (1998)Google Scholar
- 8.Durand, N., Crémilleux, B.: Ecclat: A new approach of clusters discovery in categorical data. In: 22nd SGAI International Conference on Knowledge Based Systems and Applied Artificial Intelligence, ES 2002, pp. 177–190 (2002)Google Scholar