trans_test_data, trans_test_label, _, _, _, _ = read_csv(
trans_test_path,
split_ratio=split_ratio,
header=True,
ignore_cols=["POL_ID", "DATA_MONTH", "TB_POL_BILL_MODE_CD", "MI"],
output_label="Lapse_Flag")
print(trans_train_data[0])
print("Train Data Size - ", len(trans_train_data))
print("Test Data Size - ", len(trans_test_data))
print("Splitting the data...")
# train_x = divide_batches_gen(trans_train_data, batch_size)
train_y = divide_batches(trans_train_label, batch_size)
# test_x = divide_batches_gen(trans_test_data, batch_size)
test_y = divide_batches(trans_test_label, batch_size)
train_batch_size = len(train_y)
test_batch_size = len(test_y)
logdir = "../tensorboard/transaction_model/" + datetime.datetime.now(
).strftime("%Y%m%d-%H%M%S")
saved_model_dir = "../maxlife_models/"
if not os.path.isdir(saved_model_dir):
os.mkdir(saved_model_dir)
saved_model = saved_model_dir + model_name
print("lstm data")
print(lstm_train_data[0])
print(len(lstm_train_data[0]))
# pos_weight = len(ffn_train_label) / sum(ffn_train_label)
pos_weight = np.count_nonzero(ffn_train_label == 0) / np.count_nonzero(
ffn_train_label == 1)
print("Train Data Size - ", len(ffn_train_data))
print("Creating batches...")
# train_x = divide_batches_gen(ffn_train_data, batch_size)
train_y = divide_batches(ffn_train_label, batch_size)
train_batch_size = len(train_y)
saved_model_dir = "../maxlife_models/"
if not os.path.isdir(saved_model_dir):
os.mkdir(saved_model_dir)
saved_model = saved_model_dir + model_name
ckpt = tf.train.latest_checkpoint(saved_model)
filename = ".".join([ckpt, 'meta'])
model_saver = tf.train.import_meta_graph(filename, clear_devices=True)
with tf.device("/GPU:0"):
with tf.Session() as sess:
# sess.run(init)
split_ratio = [100, 0, 0]
print("Reading the data...")
inference_data, inference_label, _, _, _, _ = read_csv(
infer_path,
split_ratio=split_ratio,
header=True,
ignore_cols=["POL_ID", "DATA_MONTH"],
output_label="Lapse_Flag")
print(inference_data[0])
print("Infer Data Size - ", len(inference_data))
print("Splitting the data...")
infer_y = divide_batches(inference_label, batch_size)
infer_batch_size = len(infer_y)
saved_model_dir = "../maxlife_models/"
if not os.path.isdir(saved_model_dir):
os.mkdir(saved_model_dir)
saved_model = saved_model_dir + model_name
ckpt = tf.train.latest_checkpoint(saved_model)
filename = ".".join([ckpt, 'meta'])
model_saver = tf.train.import_meta_graph(filename, clear_devices=True)
with tf.device("/GPU:0"):
with tf.Session() as sess:
model_saver.restore(sess, ckpt)