Filtering spam messages and mails using fuzzy C means algorithm

Advancement in computer technology has changed the world in many different ways. Communication is just a click away with the power of internet. For effective, low cost and fast communication between people email plays a very important role and thus there is a great need of email services in daily life of users. From all the transactions to business or general communication these are done through the help of emails. But often the communication is effected by the attacks on the email system which include spam mails. Spamming is the use of messaging or electronic messaging system that send huge amount of data. Spam often fills the internet with multiple copies of a message and are sent to different recipients repeatedly without their request and urges to open them. In this paper we analyze different machine learning techniques with feature selection and without feature selection algorithms and their performance to detect the best classifier for spam mail classification. First, we apply each classifier without selecting any features in order to experiment on the dataset and examine the outcome. Next, to select the desired features we apply best first feature selection algorithm and apply various algorithms for classification. We found that the accuracy has improved when we applied feature selection process in the experimentation.