Title :
On the Relationship Between Dependence Tree Classification Error and Bayes Error Rate
Author :
Balagani, Kiran S. ; Phoha, Vir V.
Author_Institution :
Louisiana Tech. Univ., Ruston
Abstract :
Wong and Poon (1989) showed that Chow and Liu´s tree dependence approximation can be derived by minimizing an upper bound of the Bayes error rate. Wong and Poon´s result was obtained by expanding the conditional entropy H(omega|X). We derive the correct expansion of H(omega|X) and present its implication.
Keywords :
Bayes methods; pattern classification; trees (mathematics); Bayes error rate; tree classification error; tree dependence approximation; Classification tree analysis; Entropy; Equations; Error analysis; Mutual information; Probability distribution; Random variables; State estimation; Upper bound; bayes error rate; classification; dependence tree approximation; entropy; mutual information; Algorithms; Artificial Intelligence; Bayes Theorem; Computer Simulation; Data Interpretation, Statistical; Models, Statistical; Pattern Recognition, Automated; Reproducibility of Results; Sensitivity and Specificity; Statistics as Topic;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
DOI :
10.1109/TPAMI.2007.1184
