The proliferation of peertopeer (P2P) systems has come with various compelling applications including file sharing based on distributed hash tables (DHTs) or other kinds of overlay networks. Searching the content of files (especially Web Search) requires multikeyword querying with scoring and ranking. Existing approaches have no way of taking into account the correlation between the keywords in the query. This paper presents our solution that incorporates the queries and behavior of the users in the P2P network such that interesting correlations can be inferred.