An Adaptive Classification approach to filter spam E-mail using Vector Space Model

The majority of previous studies of data mining have been concentrate on structured data, such as relational, transactional and data warehouse data. But, in actuality, an important section of the available information is stored in text databases, which consist of large collections of web documents from various sources, such as news articles, research papers, e-books, digital libraries, e-mails, and Web pages. Moreover, It is in increasing phase and in magnitude of terabytes of size. Among the ample of provisions of internet, e-mail facility is very useful and broadly used. Spam email is the strongly attached issue with email provision. Among various approaches developed to stop spam emails, filtering is an important and popular one. In this paper, to categorize spam and non-span email which arrives to our email id, classification method-KNNC Classification can work for better accuracy using Vector Space Model in adaptive manner. For getting accuracy in spam classification we have used two dataset- personal & Ling Spam Corpus(Lemm dataset) and apply KNNC Classification on them. We got nearly 95% of precision in spam & 86.6% of precision in nonspam and got 83% of accuracy using personal dataset and 80% using Lemm dataset using adaptive approach. We propose our own solution by reviewing the result and related work that adaptive approach using vector space model in KNNC classification method is efficiently provide better accuracy for filtering the spam mail for both smaller and larger dataset.