c11b2a426b89aad1e8da244e2a12fdec032adb9e,entity2rec/node2vec_recommender.py,,,#,70

Before Change

    print("Finished computing features after %s seconds" % (time.time() - start_time))

    evaluat.evaluate(node2vec_rec, x_test, y_test, qids_test, items_test)  // evaluates the recommender on the test set

    print("--- %s seconds ---" % (time.time() - start_time))

After Change

        args.test = "datasets/" + args.dataset + "/test.dat"

    if not args.validation:
        args.validation = "datasets/" + args.dataset + "/val.dat"

    if args.dataset == "LastFM":

        implicit = True

    else:

        implicit = args.implicit

    if args.dataset == "LibraryThing":

        threshold = 8

    else:

        threshold = args.threshold

    // initialize node2vec recommender
    node2vec_rec = Node2VecRecommender(args.dataset, p=args.p, q=args.q,
                                       walk_length=args.walk_length, num_walks=args.num_walks,
                                       dimensions=args.dimensions, window_size=args.window_size,
                                       iterations=args.iter)

    // initialize evaluator

    evaluat = Evaluator(implicit=args.implicit, threshold=args.threshold, all_unrated_items=args.all_unrated_items)

    // compute e2rec features
    x_train, y_train, qids_train, items_train, x_test, y_test, qids_test, items_test, \
    x_val, y_val, qids_val, items_val = evaluat.features(node2vec_rec, args.train, args.test,
                                                         validation=False, n_users=args.num_users,
                                                         n_jobs=args.workers, supervised=False)

    print("Finished computing features after %s seconds" % (time.time() - start_time))

    evaluat.evaluate(node2vec_rec, x_test, y_test, qids_test, items_test,
                     write_to_file="results/%s/node2vec/num%d_p%d_q%d_l%d_d%d_iter%d_winsize%d.csv"
                                   % (args.dataset, args.num_walks, args.p, args.q, args.walk_length,
                                      args.dimensions, args.iter, args.window_size),
                     baseline=True)  // evaluates the recommender on the test set

    print("--- %s seconds ---" % (time.time() - start_time))

Instances