On why discretization works for Naive-Bayes classifiers
作者： Ying Yang , Geoffrey I. Webb
DOI: 10.1007/978-3-540-24581-0_37
关键词:
摘要: We investigate why discretization can be effective in naive-Bayes learning. prove a theorem that identifies particular conditions under which will result classifiers delivering the same probability estimates as would obtained if correct density functions were employed. discuss factors might affect classification error discretization. suggest use of different techniques bias and variance generated classifiers. argue by properly managing variance, we effectively reduce error.
springer.com
core.ac.uk
sci-hub.se 参考文章(42)
Hung-Ju Huang, Tzu-Tsung Wong, Why Discretization Works for Naive Bayesian Classifiers international conference on machine learning. pp. 399- 406 ,(2000)
Ivan Bratko, Igor Kononenko, Bojan Cestnik, ASSISTANT 86: a knowledge-elicitation tool for sophisticated users EWSL'87 Proceedings of the 2nd European Conference on European Working Session on Learning. pp. 31- 45 ,(1987)
Bojan Cestnik, Estimating probabilities: a crucial task in machine learning european conference on artificial intelligence. pp. 147- 149 ,(1990)
Michael J. Pazzani, An iterative improvement approach for the discretization of numeric attributes in Bayesian classifiers knowledge discovery and data mining. pp. 228- 233 ,(1995)
João Gama, Luis Torgo, Carlos Soares, Dynamic Discretization of Continuous Attributes ibero american conference on ai. pp. 160- 169 ,(1998) , 10.1007/3-540-49795-1_14
Llanos Mora López, Inmaculada Fortes Ruiz, Rafael Morales Bueno, Francisco Triguero Ruiz, Dynamic Discretization of Continuous Values from Time Series european conference on machine learning. pp. 280- 291 ,(2000) , 10.1007/3-540-45164-1_30
Eun Bae Kong, Thomas G. Dietterich, Error-Correcting Output Coding Corrects Bias and Variance Machine Learning Proceedings 1995. pp. 313- 321 ,(1995) , 10.1016/B978-1-55860-377-6.50046-3
Ron Kohavi, David Wolpert, Bias plus variance decomposition for zero-one loss functions international conference on machine learning. pp. 275- 283 ,(1996)
Andrew McCallum, Ronald Rosenfeld, Thomas Mitchell, Andrew Y Ng, None, Improving Text Classification by Shrinkage in a Hierarchy of Classes international conference on machine learning. pp. 359- 367 ,(1998)
Kamal Nigam, Andrew McCallum, A comparison of event models for naive bayes text classification national conference on artificial intelligence. pp. 41- 48 ,(1998)