Mining relevant interval rules

TL;DR

This paper extends rule mining techniques to numerical data by developing an algorithm that extracts interval-based pattern rules, demonstrating its effectiveness on real datasets.

Contribution

It introduces a novel algorithm for mining relevant interval rules from numerical datasets, building on previous methods for rule extraction.

Findings

Successfully implemented the algorithm

Evaluated on real datasets showing promising results

Extends existing rule mining methods to numerical attributes

Abstract

This article extends the method of Garriga et al. for mining relevant rules to numerical attributes by extracting interval-based pattern rules. We propose an algorithm that extracts such rules from numerical datasets using the interval-pattern approach from Kaytoue et al. This algorithm has been implemented and evaluated on real datasets.