“…For determining, for a given user u, a given resource r, and some n ∈ N, the set (u,r) of n recommended tags we use User profile based tag recommendation [32] Tag based user profile : profile (u) = {(w1,t1), (w2,t2)..(wn,tn)} where ti ∈ T, wi is the weight of the ti, represents the importance of this tag to the user. Items are selected using balanced strategy, item tag matrix is constructed, for user u the preference relation between ti and tj for item k is calculated and ranking is done using voting, a vector is formed by the ranking tags, the profile of user u for item k is presented, ITW -Itemtagweight matrix is constructed, user similarity using Pearson correlation coefficient is calculated and most frequently used tags from the k similar user are recommended.…”