Now-a-day's numbers of documents on internet are increasing. Many traditional searching techniques may not return most accurate results in some cases. This can overcome by considering context of document i.e. content profile. In this paper, the authors have proposed a system which tries to address this issue with novel method of content profiling. Proposed system is using naive bayes text classifier and Apriori association rule to build content profile. This system accepts document from users and reply with set of similar documents based on context of document. Here they are considering document context of both input document and stored documents.