Document classification using Multinomial Naïve Bayesian Classifier
Rasineni Madana Mohana, S. Sumathi · 2014
An effective pattern discovery technique introduced antecedently that initial calculates discovered specificity patterns then evaluates the term weight consistent with the distribution of terms contained by the discovered patterns rather than the distribution in documents for discovering the misunderstanding downside. It additionally considers the influence of patterns from the negative coaching examples to search out ambiguous (noisy) patterns and check out to cut back their influence for the low-frequency downside. To beat this here a Multimodal Naive Bayesian algorithmic program is being employed for locating of patterns, since this may be the foremost acceptable one for classifying positive and negative documents. The standard results won't be in associate degree optimized manner. The prescribed methodology makes the output organized in a very specific order. The planned paper we have a tendency to use pattern (or phrase)-based approaches that perform higher as compared studies than different term-based strategies. This approach improves the accuracy of evaluating support, term weights as a result of discovered patterns are a lot of specific than whole documents.