Representing and Inferring Causalities among Classes of Multidimensional Data

When adopting Bayesian network (BN) to represent and infer probabilistic causalities among multidimensional variables, the size of the conditional probability table (CPT) associated with each variable is doomed to be large, and the causality inferences cannot be done for arbitrary evidences. In this paper, we first extend the general BN by augmenting parameters for describing causalities among classes instead of specific instances of multidimensional variables. In the extended BN, called CBN, the CPT of a variable includes the probability of each class given parent classes, while a classifier of each variable is associated to determine the class that the given evidence belongs to. Further, we give the method for approximate inferences of the CBN for arbitrary evidences. Preliminary experiments verify the feasibility of our methods.

[1]  Michael P. Wellman,et al.  Bayesian networks , 1995, CACM.

[2]  Peter Norvig,et al.  Artificial Intelligence: A Modern Approach , 1995 .

[3]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems - networks of plausible inference , 1991, Morgan Kaufmann series in representation and reasoning.

[4]  Jiawei Han,et al.  Data Mining: Concepts and Techniques , 2000 .

[5]  Sreerama K. Murthy,et al.  Automatic Construction of Decision Trees from Data: A Multi-Disciplinary Survey , 1998, Data Mining and Knowledge Discovery.

[6]  J. Ross Quinlan,et al.  Induction of Decision Trees , 1986, Machine Learning.

[7]  Sanjay Ranka,et al.  CLOUDS: A Decision Tree Classifier for Large Datasets , 1998, KDD.

[8]  Kyuseok Shim,et al.  PUBLIC: A Decision Tree Classifier that Integrates Building and Pruning , 1998, Data Mining and Knowledge Discovery.

[9]  Wray L. Buntine.  A Guide to the Literature on Learning Probabilistic Networks from Data , 1996, IEEE Trans. Knowl. Data Eng..

[10]  David A. Bell,et al.  Learning Bayesian networks from data: An information-theory based approach , 2002, Artif. Intell..

[11]  Gregory F. Cooper,et al.  The Computational Complexity of Probabilistic Inference Using Bayesian Belief Networks , 1990, Artif. Intell..

[12]  Judea Pearl,et al.  Evidential Reasoning Using Stochastic Simulation of Causal Models , 1987, Artif. Intell..

[13]  Torben Bach Pedersen,et al.  Multidimensional data modeling for complex data , 1999, Proceedings 15th International Conference on Data Engineering (Cat. No.99CB36337).