sep=",",
header=0)
# convert to implicit data and do negative sampling afterwards
data["label"] = 1
# split into train and test data based on time
train_data, test_data = split_by_ratio_chrono(data, test_size=0.2)
# specify complete columns information
sparse_col = ["sex", "occupation", "genre1", "genre2", "genre3"]
dense_col = ["age"]
user_col = ["sex", "age", "occupation"]
item_col = ["genre1", "genre2", "genre3"]
train_data, data_info = DatasetFeat.build_trainset(train_data, user_col,
item_col, sparse_col,
dense_col)
test_data = DatasetFeat.build_testset(test_data, sparse_col, dense_col)
# sample negative items for each record
train_data.build_negative_samples(data_info)
test_data.build_negative_samples(data_info)
print(data_info) # n_users: 5962, n_items: 3226, data sparsity: 0.4185 %
ytb_ranking = YouTubeRanking(task="ranking",
data_info=data_info,
embed_size=16,
n_epochs=3,
lr=1e-4,
batch_size=512,
use_bn=True,
col_names = [
"user", "item", "label", "time", "sex", "age", "occupation", "genre1",
"genre2", "genre3"
]
data = pd.read_csv("sample_data/sample_movielens_merged.csv",
sep=",",
header=0)
train, test = split_by_ratio_chrono(data, test_size=0.2)
sparse_col = ["sex", "occupation", "genre1", "genre2", "genre3"]
dense_col = ["age"]
user_col = ["sex", "age", "occupation"]
item_col = ["genre1", "genre2", "genre3"]
train_data, data_info = DatasetFeat.build_trainset(train,
user_col,
item_col,
sparse_col,
dense_col,
shuffle=False)
test_data = DatasetFeat.build_testset(test, shuffle=False)
print(data_info)
train_data.build_negative_samples(data_info,
num_neg=1,
item_gen_mode="random",
seed=2020)
test_data.build_negative_samples(data_info,
num_neg=1,
item_gen_mode="random",
seed=2222)
deepfm = DeepFM("ranking",
data_info,
sep=",",
header=0)
train_data, eval_data = split_by_ratio_chrono(data, test_size=0.2)
# specify complete columns information
sparse_col = ["sex", "occupation"]
multi_sparse_col = [["genre1", "genre2",
"genre3"]] # should be list of list
dense_col = ["age"]
user_col = ["sex", "age", "occupation"]
item_col = ["genre1", "genre2", "genre3"]
train_data, data_info = DatasetFeat.build_trainset(
train_data=train_data,
user_col=user_col,
item_col=item_col,
sparse_col=sparse_col,
dense_col=dense_col,
multi_sparse_col=multi_sparse_col,
)
eval_data = DatasetFeat.build_testset(eval_data)
print(data_info)
# do negative sampling, assume the data only contains positive feedback
train_data.build_negative_samples(data_info,
item_gen_mode="random",
num_neg=1,
seed=2020)
eval_data.build_negative_samples(data_info,
item_gen_mode="random",
num_neg=1,
seed=2222)
if __name__ == "__main__":
col_names = ["user", "item", "label", "time", "sex",
"age", "occupation", "genre1", "genre2", "genre3"]
all_data = pd.read_csv("sample_data/sample_movielens_merged.csv",
sep=",", header=0)
# use first half data as first training part
first_half_data = all_data[:(len(all_data) // 2)]
train, test = split_by_ratio_chrono(first_half_data, test_size=0.2)
sparse_col = ["sex", "occupation", "genre1", "genre2", "genre3"]
dense_col = ["age"]
user_col = ["sex", "age", "occupation"]
item_col = ["genre1", "genre2", "genre3"]
train_data, data_info = DatasetFeat.build_trainset(train, user_col, item_col,
sparse_col, dense_col,
shuffle=False)
test_data = DatasetFeat.build_testset(test, shuffle=False)
print(data_info)
train_data.build_negative_samples(data_info, num_neg=1,
item_gen_mode="random", seed=2020)
test_data.build_negative_samples(data_info, num_neg=1,
item_gen_mode="random", seed=2222)
deepfm = DeepFM("ranking", data_info, embed_size=16, n_epochs=2,
lr=1e-4, lr_decay=False, reg=None, batch_size=2048,
num_neg=1, use_bn=False, dropout_rate=None,
hidden_units="128,64,32", tf_sess_config=None)
deepfm.fit(train_data, verbose=2, shuffle=True, eval_data=test_data,
metrics=["loss", "balanced_accuracy", "roc_auc", "pr_auc",
"precision", "recall", "map", "ndcg"],
data = pd.read_csv("sample_data/sample_movielens_merged.csv",
sep=",", header=0)
train_data, eval_data = split_by_ratio_chrono(data, test_size=0.2)
# specify complete columns information
sparse_col = ["sex", "occupation"]
multi_sparse_col = [["genre1", "genre2", "genre3"]] # should be list of list
dense_col = ["age"]
user_col = ["sex", "age", "occupation"]
item_col = ["genre1", "genre2", "genre3"]
train_data, data_info = DatasetFeat.build_trainset(
train_data=train_data,
user_col=user_col,
item_col=item_col,
sparse_col=sparse_col,
dense_col=dense_col,
multi_sparse_col=multi_sparse_col,
pad_val=["missing"] # specify padding value
)
eval_data = DatasetFeat.build_testset(eval_data)
print(data_info)
# do negative sampling, assume the data only contains positive feedback
train_data.build_negative_samples(data_info, item_gen_mode="random",
num_neg=1, seed=2020)
eval_data.build_negative_samples(data_info, item_gen_mode="random",
num_neg=1, seed=2222)
deepfm = DeepFM("ranking", data_info, embed_size=16, n_epochs=2,
lr=1e-4, lr_decay=False, reg=None, batch_size=2048,
num_neg=1, use_bn=False, dropout_rate=None,