Feature extraction from discrete attributes
AuthorYıldız, Olcay Taner
MetadataShow full item record
CitationYıldız, O. T. (2010). Feature extraction from discrete attributes. Paper presented at the Proceedings - International Conference on Pattern Recognition, 3915-3918. doi:10.1109/ICPR.2010.952
In many pattern recognition applications, first decision trees are used due to their simplicity and easily interpretable nature. In this paper, we extract new features by combining k discrete attributes, where for each subset of size k of the attributes, we generate all orderings of values of those attributes exhaustively. We then apply the usual univariate decision tree classifier using these orderings as the new attributes. Our simulation results on 16 datasets from UCI repository  show that the novel decision tree classifier performs better than the proper in terms of error rate and tree complexity. The same idea can also be applied to other univariate rule learning algorithms such as C4.5Rules  and Ripper .