def test_create_bah_trader(self):
expert_a = ObscureExpert(Company.A)
expert_b = ObscureExpert(Company.B)
trader = DeepQLearningTrader(expert_a, expert_b, False, False,
'test_color', 'test_name')
self.assertIsNotNone(trader)
self.assertEqual(trader.get_color(), 'test_color')
self.assertEqual(trader.get_name(), 'test_name')
def main_orig():
"""
Code from the original main routine.
Visualizes portfolio value over the testing period for different traders.
"""
# Load stock market data for testing period
stock_market_data = StockMarketData([Company.A, Company.B],
[Period.TESTING])
# create new stock exchange with initial portfolio cash for each traders
stock_exchange = StockExchange(2000.0)
# create the traders
bah_trader = BuyAndHoldTrader('black', 'Buy and hold')
tt_trader_obscure = TrustingTrader(ObscureExpert(Company.A),
ObscureExpert(Company.B), 'green',
'Trust experts')
tt_trader2 = TrustingTrader2(ObscureExpert(Company.A),
ObscureExpert(Company.B), 'limegreen',
'Trust experts for sell only')
dql_trader = deep_q_learning_trader.DeepQLearningTrader(
ObscureExpert(Company.A),
ObscureExpert(Company.B),
True,
False,
'red',
plot_name='Deep Q-learning trader')
# run the stock exchange over the testing period, with 100 skipped trading days
stock_exchange.run(stock_market_data,
[bah_trader, tt_trader_obscure, tt_trader2, dql_trader])
# visualize the results
stock_exchange.visualize_last_run()
def main_print_model_performance():
"""
Just prints the final portfolio value and v score (return % per trading day)
for training and testing data set for different traders.
"""
stock_market_data_train = StockMarketData([Company.A, Company.B],
[Period.TRAINING])
# stock_market_data_train = stock_market_data_train.deepcopy_first_n_items(
# int(stock_market_data_train.get_row_count() / 5))
stock_market_data_test = StockMarketData([Company.A, Company.B],
[Period.TESTING])
bah_trader = BuyAndHoldTrader(name='buy and hold')
tt_trader1 = TrustingTrader(ObscureExpert(Company.A),
ObscureExpert(Company.B), 'green',
'trust experts, prefer A')
tt_trader2 = TrustingTrader2(ObscureExpert(Company.A),
ObscureExpert(Company.B), 'limegreen',
'trust experts for sell only')
dql_trader = deep_q_learning_trader.DeepQLearningTrader(
ObscureExpert(Company.A), ObscureExpert(Company.B), True, False, 'red')
all_traders = [bah_trader, tt_trader1, tt_trader2, dql_trader]
trader_names = []
for trader in all_traders:
if isinstance(trader, deep_q_learning_trader.DeepQLearningTrader
) and trader.plot_name is not None:
trader_name = trader.plot_name
else:
trader_name = trader.get_name()
trader_names.append(trader_name)
max_trader_name_length = max((len(name) for name in trader_names))
for trader, trader_name in zip(all_traders, trader_names):
is_first_line = True
for dataset, dataset_name, starting_cash in [
(stock_market_data_train, 'train', 10000.0),
(stock_market_data_test, 'test', 2000.0)
]:
stock_exchange = StockExchange(starting_cash)
stock_exchange.run(dataset, [trader])
p = stock_exchange.get_final_portfolio_value(trader)
samples = dataset.get_row_count()
v = 100.0 * (math.pow(p / starting_cash, 1 / samples) - 1.0)
if is_first_line:
header = ('{: <' + str(max_trader_name_length) +
'}').format(trader_name)
else:
header = ' ' * max_trader_name_length
header += f' ({dataset_name: <5}): '
print(f'{header}{v:.5f}% return per trading day'
f' (final portfolio value of {p:.1e})')
is_first_line = False
if isinstance(trader, BuyAndHoldTrader):
trader.reset()
# This method retrains the traders from scratch using training data from TRAINING and test data from TESTING
EPISODES = 5
if __name__ == "__main__":
# Create the training data and testing data
# Hint: You can crop the training data with training_data.deepcopy_first_n_items(n)
training_data = StockMarketData([Company.A, Company.B], [Period.TRAINING])
testing_data = StockMarketData([Company.A, Company.B], [Period.TESTING])
#training_data = training_data.deepcopy_first_n_items(6000)
# 12588
#print(f'training_data[Company.A].get_row_count(): {training_data[Company.A].get_row_count()}')
# Create the stock exchange and one traders to train the net
stock_exchange = stock_exchange.StockExchange(10000.0)
training_trader = DeepQLearningTrader(ObscureExpert(Company.A),
ObscureExpert(Company.B), False,
True)
# Save the final portfolio values per episode
final_values_training, final_values_test = [], []
for i in range(EPISODES):
logger.info(f"DQL Trader: Starting training episode {i}")
# train the net
stock_exchange.run(training_data, [training_trader])
training_trader.save_trained_model()
final_values_training.append(
stock_exchange.get_final_portfolio_value(training_trader))
pyplot.legend(trader_names)
pyplot.show()
# This main method evaluates all traders over the testing period and visualize the results.
if __name__ == "__main__":
mpl.rcParams["savefig.directory"] = os.chdir(
os.path.dirname(__file__) + '\\eval')
# Load stock market data for testing period
stock_market_data = StockMarketData([Company.A, Company.B],
[Period.TESTING])
# create new stock exchange with initial portfolio cash for each traders
stock_exchange = StockExchange(2000.0)
# create the traders
bah_trader = BuyAndHoldTrader()
tt_trader_obscure = TrustingTrader(ObscureExpert(Company.A),
ObscureExpert(Company.B), 'green',
'tt obscure')
dql_trader = deep_q_learning_trader.DeepQLearningTrader(
ObscureExpert(Company.A), ObscureExpert(Company.B), True, False, 'red')
# run the stock exchange over the testing period, with 100 skipped trading days
stock_exchange.run(stock_market_data,
[bah_trader, tt_trader_obscure, dql_trader])
# visualize the results
stock_exchange.visualize_last_run()
pyplot.figure()
trader_names = []
for trader in self.__trader_portfolios:
portfolios = self.__trader_portfolios[trader]
keys = portfolios.keys()
values = [pf.get_value(self.__complete_stock_market_data, date) for date, pf in portfolios.items()]
pyplot.plot(keys, values, label=trader.get_name(), color=trader.get_color())
trader_names.append(trader.get_name())
pyplot.legend(trader_names)
pyplot.show()
# This main method evaluates all traders over the testing period and visualize the results.
if __name__ == "__main__":
# Load stock market data for testing period
stock_market_data = StockMarketData([Company.A, Company.B], [Period.TESTING])
# create new stock exchange with initial portfolio cash for each traders
stock_exchange = StockExchange(2000.0)
# create the traders
bah_trader = BuyAndHoldTrader()
tt_trader_obscure = TrustingTrader(ObscureExpert(Company.A), ObscureExpert(Company.B), 'green', 'tt obscure')
dql_trader = deep_q_learning_trader.DeepQLearningTrader(ObscureExpert(Company.A), ObscureExpert(Company.B), True, False, 'red')
# run the stock exchange over the testing period, with 100 skipped trading days
stock_exchange.run(stock_market_data, [bah_trader, tt_trader_obscure, dql_trader])
# visualize the results
stock_exchange.visualize_last_run()
traders = dict(train=[], test=[])
def get_last_portfolio_value(phase: str, run: int) -> float:
index = min(run * EPISODES + EPISODES - 1, len(final_portfolio_values[phase]) - 1)
return final_portfolio_values[phase][index]
def get_last_v_score(phase: str, run: int) -> float:
index = min(run * EPISODES + EPISODES - 1, len(return_pct_per_day[phase]) - 1)
return return_pct_per_day[phase][index]
for run in range(TRAINING_RUNS):
# Create the stock exchange and one traders to train the net
starting_cash = dict(train=10000.0, test=2000.0)
stock_exchanges = dict(train=stock_exchange.StockExchange(starting_cash["train"]),
test=stock_exchange.StockExchange(starting_cash["test"]))
traders['train'].append(DeepQLearningTrader(ObscureExpert(Company.A), ObscureExpert(Company.B), False, True))
for episode in range(EPISODES):
# logger.info(f"DQL Trader: Starting training episode {episode}")
for phase in ['train', 'test']:
if phase == 'test':
testing_trader = DeepQLearningTrader(ObscureExpert(Company.A), ObscureExpert(Company.B), True, False)
if episode == 0:
traders[phase].append(testing_trader)
else:
traders[phase][-1] = testing_trader # replace testing trader from previous episode
trader = traders[phase][-1]
stock_exchanges[phase].run(data[phase], [trader])
if phase == 'train':
trader.save_trained_model() # required to be able to create testing trader in next iteration