Robust and cost-effective approach for discovering action rules

Kalanat, N; Shamsinejad, P; Saraee, MH

Abstract

The main goal of Knowledge Discovery in
Databases is to find interesting and usable patterns, meaningful
in their domain. Actionable Knowledge Discovery came to
existence as a direct respond to the need of finding more usable
patterns called actionable patterns. Traditional data mining
and algorithms are often confined to deliver frequent patterns
and come short for suggesting how to make these patterns
actionable. In this scenario the users are expected to act.
However, the users are not advised about what to do with
delivered patterns in order to make them usable. In this paper,
we present an automated approach to focus on not only creating
rules but also making the discovered rules actionable.
Up to now few works have been reported in this field which
lacking incomprehensibility to the user, overlooking the cost
and not providing rule generality. Here we attempt to present a
method to resolving these issues. In this paper CEARDM
method is proposed to discover cost-effective action rules from
data. These rules offer some cost-effective changes to
transferring low profitable instances to higher profitable ones.
We also propose an idea for improving in CEARDM method.

Citation

Kalanat, N., Shamsinejad, P., & Saraee, M. (2011). Robust and cost-effective approach for discovering action rules. International journal of machine learning and computing (Online), 1(4), 325-331. https://doi.org/10.7763/IJMLC.2011.V1.48

• Journal Article Type: Article
• Publication Date: Oct 1, 2011
• Deposit Date: Apr 27, 2020
• Publicly Available Date: Apr 27, 2020
• Journal: International Journal of Machine Learning and Computing
• Volume: 1
• Issue: 4
• Pages: 325-331
• DOI: https://doi.org/10.7763/IJMLC.2011.V1.48
• Publisher URL: https://doi.org/10.7763/IJMLC.2011.V1.48
• Related Public URLs: http://www.ijmlc.org/