reviews_test = [json.loads(line) for line in f] users_rev_dic = {} for rev in reviews: if not users_rev_dic.has_key(rev["user_id"]): biz_rev_dic = {rev["business_id"]: rev["stars"]} users_rev_dic[rev["user_id"]] = biz_rev_dic else: users_rev_dic[rev["user_id"]][rev["business_id"]] = rev["stars"] print 'dic built' #Build the model model = MatrixPreferenceDataModel(users_rev_dic) print 'Model built' #Build the similarity similarity = UserSimilarity(model, pearson_correlation) print 'Similarity built' #Build the User based recommender_sample.py recommender = UserBasedRecommender(model, similarity, with_preference=True) print 'Recommender built' recommender.estimate_preference("EMpFiVyiaMS58XsLZdS6vA", 'QL3vFMAsEHqfi1KGH-4igg') print 'end'