def test_case_1(self):
pipeline = Pipeline([('sparse', SparseMatrixCreator()),
('fit', NMFRecommender(n_components=2))])
preprecessed_data_dict = preprocessing(self.data_dict, True, 50, 50)
tds = TrainTestSplitter(preprecessed_data_dict, 10, 0.2)
train, test = next(tds.__iter__())
pipeline.fit(train)
y_pred = pipeline._final_estimator.predict(test)
y_true = test['ratings']['Book-Rating'].values
print(rmse(y_true, y_pred))
def test_case_2(self):
algo = SVDpp()
reader = Reader(rating_scale=(1, 10))
preprocessed_data_dict = preprocessing(self.data_dict, False, 100, 100)
preprocessed_data_dict['ratings'] = preprocessed_data_dict[
'ratings'].rename(
{
'User-ID': 'userID',
'ISBN': 'itemID',
'Book-Rating': 'rating'
},
axis='columns')
data = Dataset.load_from_df(
preprocessed_data_dict['ratings'][['userID', 'itemID', 'rating']],
reader)
my_cross_validation(algo,
data,
k=5,
threshold=7,
n_splits=5,
verbose=True)