%0 Conference Paper %1 Bringmann:EtAl:06 %A Bringmann, Bjorn %A Zimmermann, Albrecht %A Raedt, Luc De %A Nijssen, Siegfried %D 2006 %J 10th European Conference on Principles and Practice of Knowledge Discovery in Databases %K 2006 pkdd structure %P 55--66 %T Don't Be Afraid of Simpler Patterns %U http://dx.doi.org/10.1007/11871637_10 %X This paper investigates the trade-off between the expressiveness of the pattern language and the performance of the pattern miner in structured data mining. This trade-off is investigated in the context of correlated pattern mining, which is concerned with finding the k-best patterns according to a convex criterion, for the pattern languages of itemsets, multi-itemsets, sequences, trees and graphs. The criteria used in our investigation are the typical ones in data mining: computational cost and predictive accuracy and the domain is that of mining molecular graph databases. More specifically, we provide empirical answers to the following questions: how does the expressive power of the language affect the computational cost? and what is the trade-off between expressiveness of the pattern language and the predictive accuracy of the learned model? While answering the first question, we also introduce a novel stepwise approach to correlated pattern mining in which the results of mining a simpler pattern language are employed as a starting point for mining in a more complex one. This stepwise approach typically leads to significant speed-ups (up to a factor 1000) for mining graphs.

@inproceedings{Bringmann:EtAl:06, abstract = {This paper investigates the trade-off between the expressiveness of the pattern language and the performance of the pattern miner in structured data mining. This trade-off is investigated in the context of correlated pattern mining, which is concerned with finding the k-best patterns according to a convex criterion, for the pattern languages of itemsets, multi-itemsets, sequences, trees and graphs. The criteria used in our investigation are the typical ones in data mining: computational cost and predictive accuracy and the domain is that of mining molecular graph databases. More specifically, we provide empirical answers to the following questions: how does the expressive power of the language affect the computational cost? and what is the trade-off between expressiveness of the pattern language and the predictive accuracy of the learned model? While answering the first question, we also introduce a novel stepwise approach to correlated pattern mining in which the results of mining a simpler pattern language are employed as a starting point for mining in a more complex one. This stepwise approach typically leads to significant speed-ups (up to a factor 1000) for mining graphs.}, added-at = {2007-02-14T19:56:42.000+0100}, author = {Bringmann, Bjorn and Zimmermann, Albrecht and Raedt, Luc De and Nijssen, Siegfried}, biburl = {https://www.bibsonomy.org/bibtex/27a0cfc3d63bf5babcd0fb146459e1418/seandalai}, interhash = {8d51d17db7d031df34e6f7fe5e117405}, intrahash = {7a0cfc3d63bf5babcd0fb146459e1418}, journal = {10th European Conference on Principles and Practice of Knowledge Discovery in Databases}, keywords = {2006 pkdd structure}, pages = {55--66}, timestamp = {2007-02-14T19:56:42.000+0100}, title = {Don't Be Afraid of Simpler Patterns}, url = {http://dx.doi.org/10.1007/11871637_10}, year = 2006 }