Abstract Averaged onedependence estimators (AODE) is an important Bayesian learning method. However, in AODE all outputs of the superparentonedependence estimators (SPODEs) are equally considered, which may bring bad influences to the final results. In this paper, every SPODE is viewed as a generative model, and the model likelihood is used to measure its performance. Then, a new approach, named modellikelihood based superparentone dependence estimator (LODE), is proposed which integrates the SPODEs based on model likelihood. Compared with AODE, LODE significantly improves the classification performance with only a slight increase in computation.
[1] Domingos P, Pazzani M. Beyond Independence: Conditions for the Optimality of the Simple Bayesian Classifier // Proc of the 13th International Conference on Machine Learning. Bari, Italy, 1996: 105112 [2] Kononenko I. SemiNave Bayesian Classifier // Proc of the 6th European Working Session on Machine Learning. Porto, Portugal, 1991: 206219 [3] Kohavi R. Scaling Up the Accuracy of NaveBayes Classifiers: A DecisionTree Hybrid // Proc of the 2nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Menlo Park, USA, 1996: 202207 [4] Sahami M. Learning Limited Dependence Bayesian Classifiers // Proc of the 2nd International Conference on Knowledge Discovery and Data Mining. Portland, USA, 1996: 334338 [5] Friedman N, Geiger D, Goldszmidt M. Bayesian Network Classifiers. Machine Learning, 1997, 29(2): 131163 [6] Keogh E J, Pazzani M J. Learning Augmented Bayesian Classifiers: A Comparison of DistributionBased and ClassificationBased Approaches// Proc of the 7th International Workshop on Artificial Intelligence and Statistics. Ft Lauderdale, USA, 1999: 225230 [7] Zheng Z, Webb G I. Lazy Learning of Bayesian Rules. Machine Learning, 2000, 41(1): 5384 [8] Webb G I, Boughton J R, Wang Z H. Not So Nave Bayes: AggregatingOneDependenceEstimators.MachineLearning, 2005, 58(1): 524 [9] Jordan M I. Learning in Graphical Models. Cambridge, USA: MIT Press, 1998 [10] Heckerman D. A Tutorial on Learning with Bayesian Networks // Jordan M I, ed. Learning in Graphical Models. Cambridge, USA: MIT Press, 1998: 301354 [11] Garner S R. WEKA: The Waikato Environment for Knowledge Analysis // Proc of the New Zealand Computer Science Research Students Conference. Hamilton, New Zealand, 1995: 5764 [12] Blake C, Keogh E, Merz C J. UCI Repository of Machine Learning Databases [DB/OL]. [19980705]. http://www.ics.uci.edu/~mlearn/ MLRepository.html [13] Fayyad U M, Irani K B. MultiInterval Discretization of ContinuousValued Attributes for Classification Learning // Proc of the 13th International Joint Conference on Artificial Intelligence. Chambery, France, 1993: 10221027 [14]Shi Hongbo, Wang Zhihai, Huang Houkuan, et al. A Restricted DoubleLevel Bayesian Classification Model. Journal of Software, 2004, 15(2): 193199 (in Chinese) (石洪波,王志海,黄厚宽,等.一种限定性的双层贝叶斯分类模型.软件学报, 2004, 15(2): 193199)
2007, Vol. 20 
