def test_matrix_split_7(interactions_ds):
"""Test if error is thrown with an invalid value of item_test_ratio (> 1)."""
try:
matrix_split(interactions_ds, item_test_ratio=2)
except Exception as e:
assert str(
e) == 'Invalid item_test_ratio of 2: must be in the range (0, 1]'
def test_matrix_split_1(interactions_ds):
"""Test if error is thrown with an invalid value of user_test_ratio (negative)."""
try:
matrix_split(interactions_ds, user_test_ratio=-1)
except Exception as e:
assert str(
e) == 'Invalid user_test_ratio of -1: must be in the range (0, 1]'
def test_matrix_split_10(interactions_ds):
"""Test user_test_ratio = 0.7 and item_test_ratio = 0.5."""
train_ds, test_ds = matrix_split(interactions_ds,
user_test_ratio=0.7,
item_test_ratio=0.5,
seed=0)
assert [[1, 2, 3, 100, 0], [1, 4, 5, 50, 1], [1, 5, 2, 25, 2],
[2, 3, 2, 20, 4], [3, 3, 5, 250,
7]] == train_ds.values_list(to_list=True)
assert [[2, 2, 5, 100, 3], [2, 4, 2, 20, 5],
[3, 2, 1, 200, 6]] == test_ds.values_list(to_list=True)
def test_matrix_split_16(interactions_ds):
"""Test min_user_interactions = 3 so that users with #records < 3 should be removed from both sets."""
train_ds, test_ds = matrix_split(interactions_ds,
user_test_ratio=0.75,
item_test_ratio=0.75,
min_user_interactions=3,
seed=0)
assert [[1, 2, 3, 100, 0], [1, 4, 5, 50, 1], [1, 5, 2, 25, 2],
[2, 3, 2, 20, 4]] == train_ds.values_list(to_list=True)
assert [[2, 2, 5, 100, 3], [2, 4, 2, 20,
5]] == test_ds.values_list(to_list=True)
from DRecPy.Recommender.Baseline import UserKNN
from DRecPy.Dataset import InteractionDataset
from DRecPy.Evaluation.Processes import ranking_evaluation
from DRecPy.Evaluation.Splits import matrix_split
from DRecPy.Evaluation.Metrics import Precision
from DRecPy.Evaluation.Metrics import Recall
from DRecPy.Evaluation.Metrics import NDCG
ds = InteractionDataset('./cheRM_total.csv',
columns=['user', 'item', 'interaction'],
verbose=False)
ds_train, ds_test = matrix_split(ds,
min_user_interactions=20,
user_test_ratio=0.2,
item_test_ratio=0.2,
seed=25,
verbose=False)
# cosine sim
knn = UserKNN(k=10,
m=0,
sim_metric='cosine_cf',
shrinkage=None,
seed=25,
use_averages=False,
verbose=True)
knn.fit(ds_train)
evaluation = ranking_evaluation(knn,
ds_test,
from DRecPy.Recommender.Baseline import UserKNN
from DRecPy.Dataset import get_full_dataset
from DRecPy.Evaluation.Processes import ranking_evaluation
from DRecPy.Evaluation.Splits import matrix_split
from DRecPy.Evaluation.Metrics import Precision
from DRecPy.Evaluation.Metrics import Recall
from DRecPy.Evaluation.Metrics import NDCG
ds = get_full_dataset('ml-100k')
ds_train, ds_test = matrix_split(ds, user_test_ratio=0.2, item_test_ratio=0.2, seed=0, verbose=False)
# cosine sim
knn = UserKNN(k=10, m=0, sim_metric='cosine_cf', shrinkage=None, seed=15, use_averages=False, verbose=True)
knn.fit(ds_train)
evaluation = ranking_evaluation(knn, ds_test, interaction_threshold=5, k=list(range(1, 11)),
generate_negative_pairs=False, n_pos_interactions=None,
n_neg_interactions=None, seed=15, verbose=True,
metrics=[Precision(), Recall(), NDCG()])
print('cosine sim', evaluation)
# jaccard sim
knn = UserKNN(k=10, m=0, sim_metric='jaccard', shrinkage=None, seed=15, use_averages=False, verbose=True)
knn.fit(ds_train)
evaluation = ranking_evaluation(knn, ds_test, interaction_threshold=5, k=list(range(1, 11)),
generate_negative_pairs=False, n_pos_interactions=None,
n_neg_interactions=None, seed=15, verbose=True,
metrics=[Precision(), Recall(), NDCG()])
print('jaccard sim', evaluation)
def test_matrix_split_8(interactions_ds):
"""Test if error is thrown with an invalid value of max_concurrent_threads (zero)."""
try:
matrix_split(interactions_ds, max_concurrent_threads=0)
except Exception as e:
assert str(e) == 'The value of max_concurrent_threads (0) must be > 0.'
