Spam filtering using hybrid local-global Naive Bayes classifier

This paper propose a novel learning framework for classification of messages into spam and legit. We introduce a classification method based on feature space segmentation. Naive Bayes (NB) model is a statistical filtering process which uses previously gathered knowledge. Instead of using a single classifier, we propose the use of local and global classifier, based on Bayesian hierarchal framework. This helps in achieving multi-task learning, as simultaneous extraction of knowledge can be achieved while achieving classification accuracy. Knowledge among different task can be shared while learning for task specific.