Abstract

Discretization is a popular approach for handling numeric attributes in machine learning. The attributes in the datasets are both nominal and continuous. Most of the Classifiers are capable to be applied on discretized data. Hence, pre-processing of continuous data for converting them into discretized data is a necessary step before being used for the Classification Rule Mining approaches. Recently developed Associative Classifiers like CBA, CMAR and CPAR are almost equal in accuracy and have outperformed traditional classifiers. The distribution of continuous data into discrete ranges may affect the accuracy of classification. This work provides a comparative study of few discretization methods with these new classifiers. The target is to find some suitable discretization methods that are more suitable with these associative classifiers.

References

• Alcala-Fdez, J., Sanchez, L., et al. 2009. KEEL: A software tool to assess evolutionary algorithms for data mining problems. Soft Computing - A Fusion of Foundations, Methodologies and Applications: Springer Berlin / Heidelberg, pp.307-318. Available from anonymous ftp.
• Ching, J. Y., Wong Andrew, K. C., and Chan, Keith K. C. 1995. Inductive Learning from Continuous and Mixed-Mode Data. IEEE Transactions on Pattern Analysis and Machine Intelligence
• Dougherty, J., Kohavi, R. and Sahami, N. 1995 Supervised and Unsupervised Discretization of continuous Features. Machine Learning, 14th IJCAI, 1995, 108-121.
• Fayyad, U. M. and Irani, K. B. 1993. Multi-Interval Discretization of Continuous Valued Attributes for Classification Learning. 13th IJCAI, vol. 2., Chambery, France, 28.8.-2.9.93, Morgan Kaufmann, 1022–1027
• R. Giraldez, J. S., Aguilar-ruiz, et al. 2002. Discretization Oriented to Decision Rules Generation. Frontiers in Artificial Intelligence and Applications.
• Kerber, R. 1992. ChiMerge: Discretization of Numeric Attributes. In proceedings of tenth National Conference on Artificial Intelligence. 123-128
• Liu, B., Hsu, W., and Ma, Y. 1998. Integrating Classification and Association Rule Mining. In proceedings of the KDD, 1997, New York, 80-86, 1998
• Perner, P., Trautzsch, S. 1998. Multi-Interval Discretization Methods for Decision Tree Learning. LNCS 1451, Springer Verlag, 475-482
• Quinlan, J. R. 1986. Induction of Decision Trees. Machine Learning, 1: 81-106
• Quinlan, J. R., Cameron-Jones, R. M. 1993. FOIL: A midterm report. In proceedings of European Conference on Machine Learning, Vienna, Austria
• Thabtah, F., Cowling, P. and Peng, Y. 2005. A Study of Predictive Accuracy for Four Associative Classifiers. Journal Of Digital Information Management
• Thabtah, F., Cowling, P. and Peng, Y. 2006. Multiple Labels Associative Classification. Knowledge and Information Systems. Vol. 9, No. 1. 109-129
• Wenmin, L., Jiawei, H. and Pei, J. 2001. CMAR: Accurate and Efficient Classification Based on Multiple Class-Association Rules. In proceedings of ICDM. 369-376
• Xiaoxin, Y. and Han, J. 2003. CPAR: Classification based on Predictive Association Rules. In proceedings of SIAM International Conference on Data Mining, San Fransisco, CA. 331-335.