Knowledge discovery by inspection
Decision Support Systems - Special issue: knowledge discovery and its applications to business decision making
Attribute-value distribution as a technique for increasing the efficiency of data mining
Knowledge discovery and data mining
Machine Learning
A Strategy for Increasing the Efficiency of Rule Discovery in Data Mining
IDA '97 Proceedings of the Second International Symposium on Advances in Intelligent Data Analysis, Reasoning about Data
Hi-index | 0.00 |
One strategy for increasing the efficiency of rule discoveryin data mining is to target a restricted class of rules, such asexact or almost exact rules, rules with a limited number ofconditions, or rules in which each condition, on its own, eliminatesa competing outcome class. An algorithm is presented for thediscovery of rules in which each condition is a distinctive featureof the outcome class on its right-hand side in the subset of the dataset defined by the conditions, if any, which precede it. Such a ruleis said to be characteristic for the outcome class. A feature isdefined as distinctive for an outcome class if it maximises awell-known measure of rule interest or is unique to the outcome classin the data set. In the special case of data mining which ariseswhen each outcome class is represented by a single instance in thedata set, a feature of an object is shown to be distinctive if andonly if no other feature is shared by fewer objects in the dataset.