def test_random_rec_from_candidates():
algo = basic.Random()
items = lktu.ml_test.ratings['item'].unique()
users = lktu.ml_test.ratings['user'].unique()
user1, user2 = np.random.choice(users, size=2, replace=False)
algo.fit(lktu.ml_test.ratings)
# recommend from candidates
candidates = np.random.choice(items, 500, replace=False)
recs1 = algo.recommend(user1, 100, candidates=candidates)
recs2 = algo.recommend(user2, 100, candidates=candidates)
assert len(recs1) == 100
assert len(recs2) == 100
def test_random_derive_seed():
algo = basic.Random(rng_spec='user')
model = algo.fit(lktu.ml_test.ratings)
items = lktu.ml_test.ratings['item'].unique()
nitems = len(items)
assert model is not None
recs1 = algo.recommend(2038, 100)
recs2 = algo.recommend(2028, 100)
assert len(recs1) == 100
assert len(recs2) == 100
# with very high probabilities
assert set(recs1['item']) != set(recs2['item'])
recs_all = algo.recommend(2038)
assert len(recs_all) == nitems
assert set(items) == set(recs_all['item'])
def test_random():
# test case: no seed
algo = basic.Random()
model = algo.fit(lktu.ml_test.ratings)
items = lktu.ml_test.ratings['item'].unique()
users = lktu.ml_test.ratings['user'].unique()
nitems = len(items)
nusers = len(users)
assert model is not None
recs1 = algo.recommend(2038, 100)
recs2 = algo.recommend(2028, 100)
assert len(recs1) == 100
assert len(recs2) == 100
# with very high probabilities
assert set(recs1['item']) != set(recs2['item'])
recs_all = algo.recommend(2038)
assert len(recs_all) == nitems
assert set(items) == set(recs_all['item'])
from lenskit import crossfold as xf
from lenskit.algorithms import basic, Recommender, als, item_knn, user_knn
from lenskit.algorithms import funksvd
from lenskit.algorithms.implicit import BPR
from lenskit.algorithms.hpf import HPF
from sklearn.model_selection import train_test_split
import pandas as pd
import matplotlib
NUM_OF_RECS = 10 # num of items to rec for each user
ml1m = ML1M('../datasets/ml-1m')
ratings = ml1m.ratings
random = basic.Random()
popular = basic.Popular()
item_to_item_100 = item_knn.ItemItem(100)
item_to_item_200 = item_knn.ItemItem(200)
item_to_item_500 = item_knn.ItemItem(500)
user_to_user_100 = user_knn.UserUser(100)
user_to_user_200 = user_knn.UserUser(200)
user_to_user_500 = user_knn.UserUser(500)
biased_mf_50 = als.BiasedMF(50)
biased_mf_100 = als.BiasedMF(100)
biased_mf_200 = als.BiasedMF(200)
implicit_mf_50 = als.ImplicitMF(50)
implicit_mf_100 = als.ImplicitMF(100)
implicit_mf_200 = als.ImplicitMF(200)
funk_svd_mf_50 = funksvd.FunkSVD(50)
funk_svd_mf_100 = funksvd.FunkSVD(100)