Title :
Adaptive Bayesian classification using noninformative Dirichlet priors
Author :
Lynch, Robert S., Jr. ; Willett, Peter K.
Author_Institution :
Naval Undersea Warfare Center, Newport, RI, USA
Abstract :
A model is developed to illustrate the effect that adapting correctly labeled training data with possibly incorrectly labeled data has on classification performance. The model is based on a previously developed model for mislabeled training data that uses the uniform Dirichlet distribution as a noninformative prior on the symbol probabilities of each class. Two versions of the model are developed under different a priori mislabeling assumptions for the data. In the first case, the probability of mislabeling is fixed and known, and in the second, the mislabeling is marginalized out, given it is a priori uniformly distributed from zero to one-half. A formula for the average probability of error is used to illustrate results that are plotted as a function of the quantization complexity, and for varying numbers of adapted mislabeled data. In general, it is shown that even for severe mislabeling, performance improves as more data are adapted to the training set
Keywords :
Bayes methods; adaptive systems; pattern classification; probability; signal classification; a priori mislabeling assumptions; a priori uniform distribution; adapted mislabeled data; adaptive Bayesian classification; average probability; classification performance; correctly labeled training data; incorrectly labeled data; mislabeled training data; mislabeling probability; noninformative Dirichlet priors; quantization complexity; severe mislabeling; symbol probabilities; training set; uniform Dirichlet distribution; Bayesian methods; Degradation; Quantization; Testing; Training data;
Conference_Titel :
Systems, Man, and Cybernetics, 2000 IEEE International Conference on
Conference_Location :
Nashville, TN
Print_ISBN :
0-7803-6583-6
DOI :
10.1109/ICSMC.2000.884423
