def test_btd_7(self):
# normalize_rnn_input_options, predict 1
x, y, other_data = build_training_dataset(normalize_rnn_input_options,
1,
stock_data=stock_prices)
self.assertEqual(x.shape, (3637, 10, 1))
self.assertEqual(y.shape, (3637, 1))
self.assertTrue("normalize_data" in other_data)
self.assertTrue("min" in other_data["normalize_data"])
self.assertTrue("max" in other_data["normalize_data"])
data_min = np.array([np.nanmin(prices)])
data_max = np.array([np.nanmax(prices)])
self.assertEqual(other_data["normalize_data"]["min"],
data_min.tolist())
self.assertEqual(other_data["normalize_data"]["max"],
data_max.tolist())
self.assertEqual(x[0, :, 0].tolist(), ((prices[:10] - data_min) /
(data_max - data_min)).tolist())
self.assertEqual(x[1, :, 0].tolist(), ((prices[1:11] - data_min) /
(data_max - data_min)).tolist())
self.assertEqual(x[-1, :, 0].tolist(),
((prices[-11:-1] - data_min) /
(data_max - data_min)).tolist())
self.assertEqual(y[0][0], prices[10])
self.assertEqual(y[1][0], prices[11])
self.assertEqual(y[-1][0], prices[-1])
def test_btd_4(self):
# rnn_input_options, predict 10
x, y, _ = build_training_dataset(rnn_input_options,
10,
stock_data=stock_prices)
self.assertEqual(x.shape, (3619, 10, 2))
self.assertEqual(y.shape, (3619, 10))
self.assertEqual(x[0, :, 0].tolist(), prices[9:19].tolist())
self.assertEqual(x[0, 0, 1], prices[:10].mean())
self.assertEqual(x[0, 1, 1], prices[1:11].mean())
self.assertEqual(x[0, 9, 1], prices[9:19].mean())
self.assertEqual(x[1, :, 0].tolist(), prices[10:20].tolist())
self.assertEqual(x[1, 0, 1], prices[1:11].mean())
self.assertEqual(x[1, 1, 1], prices[2:12].mean())
self.assertEqual(x[1, 9, 1], prices[10:20].mean())
self.assertEqual(x[-1, :, 0].tolist(), prices[-20:-10].tolist())
self.assertEqual(x[-1, 0, 1], prices[-29:-19].mean())
self.assertEqual(x[-1, 1, 1], prices[-28:-18].mean())
self.assertEqual(x[-1, 9, 1], prices[-20:-10].mean())
self.assertEqual(y[0].tolist(), prices[19:29].tolist())
self.assertEqual(y[-1].tolist(), prices[-10:].tolist())
def test_btd_1(self):
# input_options, predict 1
x, y, _ = build_training_dataset(input_options,
1,
stock_data=stock_prices)
self.assertEqual(x.shape, (3637, 11))
self.assertEqual(y.shape, (3637, 1))
x_start = [
50.3228, 54.3227, 54.8694, 52.5974, 53.1641, 54.1221, 53.2393,
51.1629, 51.3435, 50.2802
]
x_end = [
1097.99, 1125.89, 1118.62, 1141.42, 1151.87, 1122.89, 1105.91,
1102.38, 1102.12, 1127.58
]
x_start.append(sum(x_start) / len(x_start))
x_end.append(sum(x_end) / len(x_end))
self.assertEqual(x[0].tolist(), x_start)
self.assertEqual(x[-1].tolist(), x_end)
self.assertEqual(y[0][0], 50.9122)
self.assertEqual(y[-1][0], 1128.63)
def test_btd_5(self):
# index_input_options, predict 10
x, y, _ = build_training_dataset(index_input_options,
10,
stock_data=stock_prices)
self.assertEqual(x.tolist(), [[i] for i in range(1, 11)])
self.assertEqual(y.tolist(), prices[-10:].tolist())
def test_btd_2(self):
# input_options, predict 10
x, y, _ = build_training_dataset(input_options,
10,
stock_data=stock_prices)
self.assertEqual(x.shape, (3628, 11))
self.assertEqual(y.shape, (3628, 10))
x_start = prices[:10].tolist()
x_start.append(prices[:10].mean())
x_end = prices[-20:-10].tolist()
x_end.append(prices[-20:-10].mean())
self.assertEqual(x[0].tolist(), x_start)
self.assertEqual(x[-1].tolist(), x_end)
self.assertEqual(y[0].tolist(), prices[10:20].tolist())
self.assertEqual(y[-1].tolist(), prices[-10:].tolist())
def train_models(train_models_data):
"""Trains models.
Args:
train_models_data: Train models data.
Format:
{
models: [
{
"model": "model type, matches MODEL in a model class",
"stockCode": "the predicting stock",
"modelOptions": "model options dict",
"inputOptions": "input options dict"
}
]
}
Refer to train_models_sample.json.
"""
if not path.isdir(SAVED_MODELS_DIR):
makedirs(SAVED_MODELS_DIR)
for train_model_data_idx, train_model_data in enumerate(train_models_data):
print("Model {}".format(train_model_data_idx + 1))
# initialize the model
if train_model_data["model"] == LinearRegression.MODEL:
model = LinearRegression(train_model_data["modelOptions"],
train_model_data["inputOptions"],
stock_code=train_model_data["stockCode"])
elif train_model_data["model"] == SupportVectorRegression.MODEL:
model = SupportVectorRegression(
train_model_data["modelOptions"],
train_model_data["inputOptions"],
stock_code=train_model_data["stockCode"])
elif train_model_data["model"] == LinearIndexRegression.MODEL:
model = LinearIndexRegression(train_model_data["modelOptions"],
train_model_data["inputOptions"],
train_model_data["stock_code"])
elif train_model_data["model"] == SupportVectorIndexRegression.MODEL:
model = SupportVectorIndexRegression(
train_model_data["modelOptions"],
train_model_data["inputOptions"],
train_model_data["stock_code"])
elif train_model_data["model"] == DenseNeuralNetwork.MODEL:
model = DenseNeuralNetwork(
train_model_data["modelOptions"],
train_model_data["inputOptions"],
stock_code=train_model_data["stockCode"])
# prepare the data
x, y, other_data = build_training_dataset(
train_model_data["inputOptions"], model.model_options["predict_n"])
if train_model_data["model"] in [
LinearRegression.MODEL, SupportVectorRegression.MODEL,
DenseNeuralNetwork.MODEL
]:
# get the training set
x = x[:-100]
y = y[:-100]
if "normalize" in train_model_data["inputOptions"]:
model.input_options["normalize_data"] = other_data[
"normalize_data"]
# train the model
model.train(x, y)
# save the model
model.save(SAVED_MODELS_DIR_MAP[train_model_data["model"]])