论文信息 - Improved Use of Continuous Attributes in C4.5

Improved Use of Continuous Attributes in C4.5

A reported weakness of C4.5 in domains with continuous attributes is addressed by modifying the formation and evaluation of tests on continuous attributes. An MDL-inspired penalty is applied to such tests, eliminating some of them from consideration and altering the relative desirability of all tests. Empirical trials show that the modifications lead to smaller decision trees with higher predictive accuracies. Results also confirm that a new version of C4.5 incorporating these changes is superior to recent approaches that use global discretization and that construct small trees with multi-interval splits.

J. Ross Quinlan | J. R. Quinlan

[1] Ronald L. Rivest,et al. Inferring Decision Trees Using the Minimum Description Length Principle , 1989, Inf. Comput..

[2] Peter Auer,et al. Theory and Applications of Agnostic PAC-Learning with Small Decision Trees , 1995, ICML.

[3] Leo Breiman,et al. Bagging Predictors , 1996, Machine Learning.

[4] J. Rissanen. A UNIVERSAL PRIOR FOR INTEGERS AND ESTIMATION BY MINIMUM DESCRIPTION LENGTH , 1983 .

[5] Jason Catlett,et al. On Changing Continuous Attributes into Ordered Discrete Attributes , 1991, EWSL.

[6] J. Ross Quinlan,et al. C4.5: Programs for Machine Learning , 1992 .

[7] Usama M. Fayyad,et al. Multi-Interval Discretization of Continuous-Valued Attributes for Classification Learning , 1993, IJCAI.

[8] Yoav Freund,et al. A decision-theoretic generalization of on-line learning and an application to boosting , 1995, EuroCOLT.

[9] Philip J. Stone,et al. Experiments in induction , 1966 .

[10] Ron Kohavi,et al. Automatic Parameter Selection by Minimizing Estimated Error , 1995, ICML.

[11] Ron Kohavi,et al. Irrelevant Features and the Subset Selection Problem , 1994, ICML.

[12] Ron Kohavi,et al. Supervised and Unsupervised Discretization of Continuous Features , 1995, ICML.