Universitat Oberta de Catalunya
Tree classification and the frequent tree discovery task have been important tasks over the last decade. Now-a-days, they are becoming harder, as the size of the trees datasets is increasing and the authors cannot assume that data has been generated from a static distribution. They propose a new method to classify trees, using closed and maximal frequent trees. Closed trees maintain the same information as frequent trees using less space and maximal trees maintain approximate information. They use them to reduce the number of classification features. They present a new framework for data stream tree classification.