An expected utility-based approach for mining action rules

One of the central issues in data mining community is to make the mined patterns actionable. Action rules are those actionable patterns, which provide hints to a user what actions (i.e., changes within some values of flexible attributes) should be taken to reclassify some objects from an undesired decision class to a desired one. Both changing the value of a flexible attribute and the corresponding change of the value of a decision attribute may incur cost (negative utility) or bring benefit (positive utility) for the user. Obviously, the user is more interested in the rules which are expected to bring higher utility. In this paper, we formally define the expected utility of an action rule for measuring its interestingness. Our definitions explicitly state the problem of mining action rules as a search problem in a framework of support and expected utility. We also propose an effective algorithm for mining action rules with higher expected utilities. Our experiment shows the usefulness of the proposed approach.

[1]  Zbigniew W. Ras,et al.  Action-Rules: How to Increase Profit of a Company , 2000, PKDD.

[2]  Ramakrishnan Srikant,et al.  Fast algorithms for mining association rules , 1998, VLDB 1998.

[3]  Harleen Kaur,et al.  Actionable Rules: Issues and New Directions , 2007, WEC.

[4]  Zbigniew W. Ras,et al.  ARAS: Action Rules Discovery Based on Agglomerative Strategy , 2007, MCD.

[5]  Zbigniew W. Ras,et al.  Association Action Rules , 2008, 2008 IEEE International Conference on Data Mining Workshops.

[6]  Jian Pei,et al.  Mining Frequent Patterns without Candidate Generation: A Frequent-Pattern Tree Approach , 2006, Sixth IEEE International Conference on Data Mining - Workshops (ICDMW'06).

[7]  Zdzislaw Pawlak,et al.  Information systems theoretical foundations , 1981, Inf. Syst..

[8]  Zbigniew W. Ras,et al.  In Search for Action Rules of the Lowest Cost , 2004, MSRAS.

[9]  Wenji Mao,et al.  Mining actionable behavioral rules from group data , 2011, Proceedings of 2011 IEEE International Conference on Intelligence and Security Informatics.

[10]  Zbigniew W. Ras,et al.  Action Rules Discovery, a New Simplified Strategy , 2006, ISMIS.

[11]  Ramakrishnan Srikant,et al.  Fast Algorithms for Mining Association Rules in Large Databases , 1994, VLDB.

[12]  Zbigniew W. Ras,et al.  Action rule discovery from incomplete data , 2010, Knowledge and Information Systems.

[13]  Zbigniew W. Ras,et al.  Constraint Based Action Rule Discovery with Single Classification Rules , 2007, RSFDGrC.

[14]  Zbigniew W. Ras,et al.  Discovering Extended Action-Rules (System DEAR) , 2003, IIS.

[15]  Angelina A. Tzacheva,et al.  Action rules mining: Research Articles , 2005 .

[16]  Zbigniew W. Ras,et al.  Mining E-Action Rules, System DEAR , 2008, Data Mining: Foundations and Practice.

[17]  Zengyou He,et al.  Mining action rules from scratch , 2005, Expert Syst. Appl..

[18]  Qiang Yang,et al.  Mining case bases for action recommendation , 2002, 2002 IEEE International Conference on Data Mining, 2002. Proceedings..

[19]  Abraham Silberschatz,et al.  What Makes Patterns Interesting in Knowledge Discovery Systems , 1996, IEEE Trans. Knowl. Data Eng..

[20]  Zbigniew W. Ras,et al.  Mining for interesting action rules , 2005, IEEE/WIC/ACM International Conference on Intelligent Agent Technology.

[21]  Angelina A. Tzacheva,et al.  Tree-based Construction of Low-cost Action Rules , 2008 .

[22]  Qiang Yang,et al.  Mining optimal actions for profitable CRM , 2002, 2002 IEEE International Conference on Data Mining, 2002. Proceedings..