Section: New Software and Platforms

Relevant Interval Rules Miner

Keywords: Association rule - Pattern discovery - Formal concept analysis

Functional Description: This software extracts relevant rules from a dataset of labeled numerical attributes (tabular datasets). A rule is an interval-based pattern associated to a predicted label. The tool extracts a subset of rules based on the accuracy and relevance criteria where most of the algorithms are simply based on accuracy. This allow us to extract the best rules that capture the data behavior.

• Participants: René Quiniou, Véronique Masson and Thomas Guyet

• Contact: Thomas Guyet

• Publication: Mining relevant interval rules