def load_dataset(self):
print('loading ratings from path: ' + self.rates_file_dir)
train_data_frame, test_data_frame = self.__load_rate(
self.rates_file_dir, self.split_ratio)
self.user_ids = self.data['User ID'].drop_duplicates().tolist()
self.item_ids = self.data['Item ID'].drop_duplicates().tolist()
if self.type == DataType.ndarray:
self.train_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_trainset',
len(self.user_ids), len(self.item_ids))
self.test_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_testset',
len(self.user_ids), len(self.item_ids))
elif self.type == DataType.dataframe:
self.train_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
elif self.type == DataType.dictionary:
self.train_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
self.train_dataset.load_dataset(train_data_frame)
self.test_dataset.load_dataset(test_data_frame)
print('Done.')
def __init__(self,
name='MovieLensDataset100K',
type=DataType.ndarray,
split_ratio=0.8,
dataset_path='../data/MovieLens/ml-100k'):
self.name = name
self.type = type
self.users_info_path = os.path.join(dataset_path, 'u.user')
self.items_info_path = os.path.join(dataset_path, 'u.item')
self.rates_file_path = os.path.join(dataset_path, 'u.data')
self.split_ratio = split_ratio
self.users = self.__load_user_info(self.users_info_path)
self.items = self.__load_item_info(self.items_info_path)
occupation_group = self.users.groupby(by='Occupation')
occupation_list = list(occupation_group.groups.keys())
self.occupation2id = {}
for i, occupation in enumerate(occupation_list):
self.occupation2id[occupation] = i
self.gender2id = {'M': 0, 'F': 1}
self.user_ids = self.users['ID'].drop_duplicates().tolist()
self.item_ids = self.items['ID'].drop_duplicates().tolist()
self.user_num = len(self.user_ids)
self.item_num = len(self.item_ids)
self.user_occupation_num = len(self.occupation2id)
self.users = self.users.set_index('ID').sort_index()
self.items = self.items.set_index('ID').sort_index()
if self.type == DataType.ndarray:
self.train_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_trainset',
len(self.user_ids), len(self.item_ids))
self.test_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_testset',
len(self.user_ids), len(self.item_ids))
elif self.type == DataType.dataframe:
self.train_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
elif self.type == DataType.dictionary:
self.train_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
def __init__(self,
name='BookCrossingDataset',
type=DataType.dictionary,
split_ratio=0.8,
dataset_path='G:/dataset/RS/Book-crossing/BX-CSV-Dump'):
self.name = name
self.type = type
self.users_info_path = os.path.join(dataset_path, 'BX-Users.csv')
self.items_info_path = os.path.join(dataset_path, 'BX-Books.csv')
self.rates_file_path = os.path.join(dataset_path,
'BX-Book-Ratings.csv')
self.split_ratio = split_ratio
self.users = self.__load_user_info(self.users_info_path)
self.items = self.__load_item_info(self.items_info_path)
self.data = pd.read_csv(self.rates_file_path,
sep=';',
skiprows=[0],
names=['User ID', 'Item ID', 'Rating'],
dtype={
'User ID': str,
'Item ID': str,
'Rating': np.float16
},
encoding='latin1')
self.user_ids = self.data['User ID'].drop_duplicates().tolist()
self.item_ids = self.data['Item ID'].drop_duplicates().tolist()
if self.type == DataType.ndarray:
self.train_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_trainset',
len(self.user_ids), len(self.item_ids))
self.test_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_testset',
len(self.user_ids), len(self.item_ids))
elif self.type == DataType.dataframe:
self.train_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
elif self.type == DataType.dictionary:
self.train_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
def __init__(self,
name='MovieLensDataset100K',
type=DataType.ndarray,
split_ratio=0.8,
dataset_path='G:/dataset/RS/MovieLens/ml-100k'):
self.name = name
self.type = type
self.users_info_path = os.path.join(dataset_path, 'u.user')
self.items_info_path = os.path.join(dataset_path, 'u.item')
self.rates_file_path = os.path.join(dataset_path, 'u.data')
self.split_ratio = split_ratio
self.users = self.__load_user_info(self.users_info_path)
self.items = self.__load_item_info(self.items_info_path)
self.user_ids = self.users['ID'].drop_duplicates().tolist()
self.item_ids = self.items['ID'].drop_duplicates().tolist()
if self.type == DataType.ndarray:
self.train_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_trainset',
len(self.user_ids), len(self.item_ids))
self.test_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_testset',
len(self.user_ids), len(self.item_ids))
elif self.type == DataType.dataframe:
self.train_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
elif self.type == DataType.dictionary:
self.train_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
def __init__(self,
name='MovieLensDataset1M',
type=DataType.dictionary,
split_ratio=0.8,
dataset_path='G:/dataset/RS/MovieLens/ml-1m'):
self.name = name
self.type = type
self.users_info_path = os.path.join(dataset_path, 'users.dat')
self.items_info_path = os.path.join(dataset_path, 'movies.dat')
self.rates_file_path = os.path.join(dataset_path, 'ratings.dat')
self.split_ratio = split_ratio
self.users = self.__load_user_info(self.users_info_path)
self.items = self.__load_item_info(self.items_info_path)
self.user_ids = list(range(1, 6041))
self.item_ids = list(range(1, 3953))
if self.type == DataType.ndarray:
self.train_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_trainset',
len(self.user_ids), len(self.item_ids))
self.test_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_testset',
len(self.user_ids), len(self.item_ids))
elif self.type == DataType.dataframe:
self.train_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
elif self.type == DataType.dictionary:
self.train_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
def __init__(self,
name='MovieLensDataset25M',
type=DataType.dictionary,
split_ratio=0.8,
dataset_path='G:/dataset/RS/MovieLens/ml-25m'):
self.name = name
self.type = type
self.items_info_path = os.path.join(dataset_path, 'movies.csv')
self.rates_file_path = os.path.join(dataset_path, 'ratings.csv')
self.user_num = 162541
self.split_ratio = split_ratio
self.items = self.__load_item_info(self.items_info_path)
self.user_ids = list(range(1, self.user_num + 1))
self.item_ids = self.items['ID'].drop_duplicates().tolist()
if self.type == DataType.ndarray:
self.train_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_trainset',
len(self.user_ids), len(self.item_ids))
self.test_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_testset',
len(self.user_ids), len(self.item_ids))
elif self.type == DataType.dataframe:
self.train_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
elif self.type == DataType.dictionary:
self.train_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
def __init__(self,
name='MovieLensDataset1M',
type=DataType.dictionary,
split_ratio=0.8,
dataset_path='../data/MovieLens/ml-1m'):
self.name = name
self.type = type
self.users_info_path = os.path.join(dataset_path, 'users.dat')
self.items_info_path = os.path.join(dataset_path, 'movies.dat')
self.rates_file_path = os.path.join(dataset_path, 'ratings.dat')
self.split_ratio = split_ratio
self.users = self.__load_user_info(self.users_info_path)
self.items = self.__load_item_info(self.items_info_path)
self.gender2id = {'M': 0, 'F': 1}
self.category2id = {
'Action': 0,
'Adventure': 1,
'Animation': 2,
"Children's": 3,
'Comedy': 4,
'Crime': 5,
'Documentary': 6,
'Drama': 7,
'Fantasy': 8,
'Film-Noir': 9,
'Horror': 10,
'Musical': 11,
'Mystery': 12,
'Romance': 13,
'Sci-Fi': 14,
'Thriller': 15,
'War': 16,
'Western': 17
}
self.user_ids = self.users['ID'].drop_duplicates().tolist()
self.item_ids = self.items['ID'].drop_duplicates().tolist()
self.user_occupation_num = 21
self.users = self.users.set_index('ID').sort_index()
self.items = self.items.set_index('ID').sort_index()
self.user_ids = list(range(1, 6041))
self.item_ids = list(range(1, 3953))
self.user_num = len(self.user_ids)
self.item_num = len(self.item_ids)
if self.type == DataType.ndarray:
self.train_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_trainset',
len(self.user_ids), len(self.item_ids))
self.test_dataset = RatingNDArray(
self.name + '_' + self.type.name + '_testset',
len(self.user_ids), len(self.item_ids))
elif self.type == DataType.dataframe:
self.train_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDataFrame(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)
elif self.type == DataType.dictionary:
self.train_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_trainset', self.user_ids,
self.item_ids)
self.test_dataset = RatingDictionary(
self.name + '_' + self.type.name + '_testset', self.user_ids,
self.item_ids)