A Survey Paper On Naive Bayes Classifier For Multi-Feature Based Text Mining

Text mining is variance of a field called data mining. To make unstructured data workable by the computer Text mining is used which is also referred as “Text Analytics”. Text categorization, also called as topic spotting is the task of automatically classifies a set of documents into groups from a predefined set. Text classification is an essential application and research topic because of increase in digital documents. Today, text classification is an important due to the very large amount of text documents that we have to deal with daily. In past various methods have been introduced to text classification such as comparison between multi-variate Bernoulli model and multinomial model, Support Vector Machines (SVMs), using the X square test, Information Gain(IG) and the Maximum Entropy (ME) approach. In this paper we aim to classify the documents using the naive Bayes classifier and the Maximum Entropy classification model. Also we use Support Vector Machines (SVMs) and using the X square test for effective text categorization. The naive Bayes classifier and the Maximum Entropy classification model is probabilistic classifier. Support Vector Machines are a group of supervised learning methods that can be applied to classification and the X square test is used to select word features. Naive Bayes classifier is a standard method for text categorization, the difficulty in judging documents as belonging to either one category or the other with word frequencies as the features.