display_step = args.display_step
batch_size = args.batch_size
train_data, test_data, n_user, n_item = load_data_neg(test_size=0.2,
sep="\t")
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
with tf.Session(config=config) as sess:
model = None
# Model selection
if args.model == "CDAE":
train_data, test_data, n_user, n_item = load_data_all(
test_size=0.2, sep="\t")
model = ICDAE(sess, n_user, n_item)
if args.model == "CML":
model = CML(sess, n_user, n_item)
if args.model == "LRML":
model = LRML(sess, n_user, n_item)
if args.model == "BPRMF":
model = BPRMF(sess, n_user, n_item)
if args.model == "NeuMF":
model = NeuMF(sess, n_user, n_item)
if args.model == "GMF":
model = GMF(sess, n_user, n_item)
if args.model == "MLP":
model = MLP(sess, n_user, n_item)
if args.model == "JRL":
model = JRL(sess, n_user, n_item)
# build and execute the model
for u in range(num_users):
neg_items[u] = list(all_items - set(train_data.getrow(u).nonzero()[1]))
train_list.append(list(train_data.getrow(u).toarray()[0]))
train_data = train_list
# train model
model = None
kwargs = {'epoch': args.num_epochs,
'T': args.display_step,
'learning_rate': args.learning_rate,
'reg_rate': args.reg_rate,
'male_weight': args.male_weight,
'log_file': log_file}
if args.model == 'BPRMF':
model = BPRMF(sess, num_users, num_items, **kwargs)
if args.model == 'CDAE':
model = ICDAE(sess, num_users, num_items, **kwargs)
if args.model == 'CML':
model = CML(sess, num_users, num_items, **kwargs)
if args.model == 'GMF':
model = GMF(sess, num_users, num_items, **kwargs)
if args.model == 'JRL':
model = JRL(sess, num_users, num_items, **kwargs)
if args.model == 'LRML':
model = LRML(sess, num_users, num_items, **kwargs)
if args.model == 'MLP':
model = MLP(sess, num_users, num_items, **kwargs)
if args.model == 'NeuMF':
model = NeuMF(sess, num_users, num_items, **kwargs)
if model is None:
exit()
model.build_network(num_factor=args.num_factors)
display_step = args.display_step
batch_size = args.batch_size
# train_data, test_data, n_user, n_item = load_data_neg(test_size=0.2, sep="\t")
# train_data, test_data, n_user, n_item = load_data_myneg(test_size=0.2, sep=";;")
train_data, test_data, n_user, n_item, test_data_hot, test_data_long, hot_item, long_item = load_data_myneg_tail(
test_size=0.2, sep=";;")
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
with tf.Session(config=config) as sess:
model = None
# Model selection
if args.model == "CDAE":
train_data, test_data, n_user, n_item = load_data_all(
test_size=0.2, sep="\t")
model = ICDAE(sess, n_user, n_item)
if args.model == "CML":
model = CML(sess, n_user, n_item)
if args.model == "LRML":
model = LRML(sess, n_user, n_item)
if args.model == "BPRMF":
model = BPRMF(sess, n_user, n_item)
if args.model == "NeuMF":
model = NeuMF(sess, n_user, n_item)
if args.model == "GMF":
model = GMF(sess, n_user, n_item)
if args.model == "MLP":
model = MLP(sess, n_user, n_item)
if args.model == "JRL":
model = JRL(sess, n_user, n_item)
if args.model == "NeuMF_my":