Decision tree pruning: biased or optimal?
作者： Sholom M. Weiss , Nitin Indurkhya
DOI:
关键词:
摘要: We evaluate the performance of weakest-link pruning decision trees using cross-validation. This technique maps tree into a problem selection: Find best (i.e. right-sized) tree, from set ranging in size unpruned to null tree. For samples with at least 200 cases, extensive empirical evidence supports following conclusions relative (a) 10-fold cross-validation is nearly unbiased; (b) not covering highly biased; (c) consistent optimal selection for large sample sizes and (d) accuracy by largely dependent on size, irrespective population distribution.
参考文章(18)
William W. Cohen, Efficient Pruning Methods for Separate-and-Conquer Rule Learning Systems. international joint conference on artificial intelligence. pp. 988- 994 ,(1993)
Sholom M. Weiss, Nitin Indurkhya, Small Sample Decision Tree Pruning Machine Learning Proceedings 1994. pp. 335- 342 ,(1994) , 10.1016/B978-1-55860-335-6.50048-9
Cullen Schaffer, Deconstructing the Digit Recognition Problem international conference on machine learning. pp. 394- 399 ,(1992) , 10.1016/B978-1-55860-247-2.50056-5
Richard A Olshen, Charles J Stone, Leo Breiman, Jerome H Friedman, Classification and regression trees ,(1983)
Paul Everett Utgoff, Shift of bias for inductive concept learning Rutgers University. ,(1984)
Bojan Cestnik, Ivan Bratko, On estimating probabilities in tree pruning Lecture Notes in Computer Science. pp. 138- 150 ,(1991) , 10.1007/BFB0017010
Cullen Schaffer, Sparse data and the effect of overfitting avoidance in decision tree induction national conference on artificial intelligence. pp. 147- 152 ,(1992)
David H. Wolpert, On Overfitting Avoidance as Bias Research Papers in Economics. ,(1993)
S.M. Weiss, N. Indurkhya, Optimized rule induction IEEE Intelligent Systems. ,vol. 8, pp. 61- 69 ,(1993) , 10.1109/64.248354
Cullen Schaffer, Overfitting Avoidance as Bias Machine Learning. ,vol. 10, pp. 153- 178 ,(1993) , 10.1023/A:1022653209073