Abstract:Naive Bayes has been widely used in data mining as a simple and effective classification algorithm. Since its conditional independence assumption is rarely true, numerous algorithms have been proposed to improve naive Bayes, among which tree augmented naive Bayes (TAN) [3] achieves a significant improvement in term of classification accuracy, while maintaining efficiency and model simplicity. In many real-world data mining applications, however, an accurate ranking is more desirable than a classification. Thus it is interesting whether TAN also achieves significant improvement in term of ranking, measured by AUC(the area under the Receiver Operating Characteristics curve) [8,1]. Unfortunately, our experiments show that TAN performs even worse than naive Bayes in ranking. Responding to this fact, we present a novel learning algorithm, called forest augmented naive Bayes (FAN), by modifying the traditional TAN learning algorithm. We experimentally test our algorithm on all the 36 data sets recommended by Weka [12], and compare it to naive Bayes, SBC [6], TAN [3], and C4.4 [10], in terms of AUC. The experimental results show that our algorithm outperforms all the other algorithms significantly in yielding accurate rankings. Our work provides an effective and efficient data mining algorithm for applications in which an accurate ranking is required.

Tree-Augmented Naive Bayes Ensembles

Directed Tree Algorithm Based on TAN Classifiers

Improving Tree Augmented Naive Bayes for Class Probability Estimation

Generalized tree augmented naive bayes

Tree Augmented Naïve possibilistic network classifier.

Selective Augmented Bayesian Network Classifiers Based on Rough Set Theory

Hierarchical Dependency Constrained Tree Augmented Naive Bayes Classifiers for Hierarchical Feature Spaces

Structure Extension of Tree-Augmented Naive Bayes

Learning Robust Bayesian Network Classifiers in the Space of Markov Equivalent Classes

Learning Tree-Augmented Naive Bayesian Network by Reduced Space Requirements

A Selective Tree-Augmented Bayesian Network Classifier Based on Rough Set Theory

Learning High-Dependence Bayesian Network Classifier with Robust Topology

Improving Classification Performance By Combining Multiple Tan Classifiers

SubTree Augmented Naïve Bayesian classifier based on the fuzzy equivalence partition of attribute variables

A Smart Design for the TAN Classifier

A NEW BAYESIAN TREE CLASSIFIER AND ITS APPLICATION

A New Hierarchical Redundancy Eliminated Tree Augmented Naive Bayes Classifier for Coping with Gene Ontology-based Features

An active TAN classifier based on vote entropy-maximum entropy of QBC

Stock Price Forecast Using Tree Augmented Naïve (TAN) Bayes.

Adjusting Dependence Relations for Semi-Lazy TAN Classifiers

Learning Tree Augmented Naive Bayes for Ranking