TAN classifiers based on decomposable distributions

被引:37
作者
Cerquides, J
De Mantaras, RL
机构
[1] Univ Barcelona, Dept Matemat Aplicada & Anal, E-08007 Barcelona, Spain
[2] CSIC, Inst Invest & Intelligencia Artificial, Bellaterra 08190, Spain
关键词
Bayesian networks; Bayesian network classifiers; naive Bayes; tree augmented naive Bayes; decomposable distributions; Bayesian model averaging;
D O I
10.1007/s10994-005-0470-7
中图分类号
TP18 [人工智能理论];
学科分类号
081104 [模式识别与智能系统]; 0812 [计算机科学与技术]; 0835 [软件工程]; 1405 [智能科学与技术];
摘要
In this paper we present several Bayesian algorithms for learning Tree Augmented Naive Bayes (TAN) models. We extend the results in Meila & Jaakkola (2000a) to TANs by proving that accepting a prior decomposable distribution over TAN's, we can compute the exact Bayesian model averaging over TAN structures and parameters in polynomial time. Furthermore, we prove that the k-maximum a posteriori (MAP) TAN structures can also be computed in polynomial time. We use these results to correct minor errors in Meila & Jaakkola (2000a) and to construct several TAN based classifiers. We show that these classifiers provide consistently better predictions over Irvine datasets and artificially generated data than TAN based classifiers proposed in the literature.
引用
收藏
页码:323 / 354
页数:32
相关论文
共 14 条
[1]
[Anonymous], P 16 C UNC ART INT
[2]
CERQUIDES J, 1999, P INT C KNOWL DISC D, P292
[3]
APPROXIMATING DISCRETE PROBABILITY DISTRIBUTIONS WITH DEPENDENCE TREES [J].
CHOW, CK ;
LIU, CN .
IEEE TRANSACTIONS ON INFORMATION THEORY, 1968, 14 (03) :462-+
[4]
On the optimality of the simple Bayesian classifier under zero-one loss [J].
Domingos, P ;
Pazzani, M .
MACHINE LEARNING, 1997, 29 (2-3) :103-130
[5]
FAWCETT T, 2003, HPL20034 HP LAB PAL
[6]
Bayesian network classifiers [J].
Friedman, N ;
Geiger, D ;
Goldszmidt, M .
MACHINE LEARNING, 1997, 29 (2-3) :131-163
[7]
A simple generalisation of the area under the ROC curve for multiple class classification problems [J].
Hand, DJ ;
Till, RJ .
MACHINE LEARNING, 2001, 45 (02) :171-186
[8]
LEARNING BAYESIAN NETWORKS - THE COMBINATION OF KNOWLEDGE AND STATISTICAL-DATA [J].
HECKERMAN, D ;
GEIGER, D ;
CHICKERING, DM .
MACHINE LEARNING, 1995, 20 (03) :197-243
[9]
IDE J, 2003, GENERATION RANDOM BA
[10]
AN ALGORITHM FOR FINDING K-MINIMUM SPANNING-TREES [J].
KATOH, N ;
IBARAKI, T ;
MINE, H .
SIAM JOURNAL ON COMPUTING, 1981, 10 (02) :247-255