Efficient Implementation of Class Based Decomposition Schemes for Naive Bayes Classifier

The accuracy of a Naive Bayes Classifier over text classification tasks can be significantly improved by taking advantage of decomposition schemes such as Error-Correcting Output Codes. ECOC is the task of document categorization and it is a method for decomposing multi-way classification problems into binary classification tasks. By The Additive nature of the classifier all binary classifiers can be trained in single pass through the data, through this approach the training complexity is reduces O (n. t. g) to O ( (n+t). g).