Effect of Simpson's Paradox on Market Basket Analysis

10.29973/JCSA.200406.0007

Heather Y. Ma；Dennis K. J. Lin

Association Rule ； confidence ； improvement ； Odds Ratio Test ； support

中國統計學報

42卷2期（2004 / 06 / 01）

209 - 221

英文

Association rules discovery is an important database mining algorithm that finds interesting association or correlation relationships among a set of items. One of the well-studied problems in data mining is pruning for association rules in Market Basket Analysis. In this paper, we address the effect of Simpson's Paradox on the decision maker in Market Basket Analysis and proposed a method, called common improvement, to handle Simpson's Paradox in selecting association rules. The proposed method is suited for detecting association rules that are likely to be ignored from the existing methods.

1. Megaputer Intelligence Inc.
2. Fabris, C. C.,Freitas A. A.(1999).Discovering surprising patterns by detecting occurrences of Simpson`s paradox.Research and Development in Intelligent Systems XVI (Proc. ES99, The 19th SGES Int. Conf. on Knowledge-Based Systems and Applied Artificial Intelligence),148-160.
3. Freitas, A. A.(1998).On objective measures of rule Surprisingness. Principles of Data Mining & Knowledge Discovery.Artificial Intelligence,1510,1-9.
4. Freitas, A. A.(1997).England,University of Essex.
5. George, J. H.(1997).PhD thesis. Computer Science Department, School of Engineering, Stanford University.
6. Michalski, R. S.,Kerschberg, L.,Kaufman, K. A.,Ribeiro, J. S.(1992).Mining For Knowledge In Databases: The Inlen Architecture, Initial Implementation And First Results.Journal of Intelligent Information Systems
7. Simpson, E. H.(1951).The interpretation of interaction in contingency tables.J. Roy. Statistic. Soc. Ser. B,13,238-241.

1. 蔡崇煌,林彥廷,林正介(2021)。辛普森悖論解決之道：以新英格蘭醫學雜誌之羅格列酮統合分析為例。台灣家庭醫學雜誌，31(4)，245-257。