Catalogue of Artificial Intelligence Techniques
k-optimal pattern discovery.
Author(s): Flaminia Cavallo
K-optimal pattern discovery is a exploratory technique used in data-mining to seek a number of patterns, specified by the user, which optimize a measure of interest defined by the user as well.Usually data-mining tecnique search between the data for a single model in order to optimize an objective function. Most of the times there are several other models that are equally optimal and that could be prefered by the user. These preference could be based on critirias that are impossible to include in the objective function. To solve this problem, exploratory pattern discovery techniques seek for all these models, between which the user can select his preference. Between these tecniques, the k-optimal pattern discovery allows the user to specify the number of patterns to be returned, k , together with a set of constraints and the rule value measure of interest to optimize. The k-optimal constraint is often sufficient to enable efficient search
- Webb, G.I., Advances in Artificial Intelligence, vol. 3809/2005, Springer Berlin / Heidelberg.
- Webb G. I. and Zhang S., K-optimal rule discovery Data Mining and Knowledge Discovery (2005).