)
driver.add_testcase(
"continuous-time",
hedge_models.FeatureHedge(),
risk_measure=hedge_models.ExpectedShortfall(alpha),
normaliser=None,
feature_function="delta",
price_type="arbitrage")
driver.add_testcase(
"deep network",
hedge_models.NeuralHedge(
timesteps=timesteps * hedge_multiplier,
instrument_dim=book.instrument_dim,
internal_dim=0,
num_layers=layers,
num_units=units,
activation=activation),
risk_measure=hedge_models.ExpectedShortfall(alpha),
normaliser=preprocessing.MeanVarianceNormaliser(),
feature_function="log_martingale",
price_type="arbitrage")
driver.add_testcase(
"deep network w. PCA",
hedge_models.NeuralHedge(
timesteps=timesteps * hedge_multiplier,
instrument_dim=book.instrument_dim,
internal_dim=0,
frequency=0, # no need for frequency for non-path dependent derivatives.
init_instruments=init_instruments,
init_numeraire=init_numeraire,
book=book,
cost=1 / 100 if cost else None,
risk_neutral=True,
learning_rate=1e-1)
driver.verbose = 2
risklevels = [0.05, 0.5, 0.95] if not cost else [0.95]
for alpha in risklevels:
driver.add_testcase(f"deep network {alpha}",
hedge_models.NeuralHedge(
timesteps=timesteps,
instrument_dim=book.instrument_dim,
internal_dim=0,
num_layers=4,
num_units=5,
activation=tf.keras.activations.softplus),
risk_measure=hedge_models.ExpectedShortfall(alpha),
normaliser=preprocessing.MeanVarianceNormaliser(),
feature_function="log_martingale",
price_type="arbitrage")
if driver.cost is not None or not driver.risk_neutral:
driver.add_liability_free(
hedge_models.LinearFeatureHedge(
timesteps=timesteps,
instrument_dim=book.instrument_dim,
mappings=[approximators.IdentityFeatureMap] \
* (1 + (driver.cost is not None))),
normaliser=None,
feature_function=lazy_feature_function,
price_type="arbitrage")
driver.add_testcase("jump continuous-time",
hedge_models.FeatureHedge(),
risk_measure=risk_measure(),
normaliser=None,
feature_function=jump_feature_function,
price_type="arbitrage")
driver.add_testcase("deep network",
hedge_models.NeuralHedge(
timesteps=timesteps,
instrument_dim=tradebook.instrument_dim,
internal_dim=0,
num_layers=layers,
num_units=units,
activation=activation),
risk_measure=risk_measure(),
normaliser=preprocessing.MeanVarianceNormaliser(),
feature_function="log_martingale",
price_type="arbitrage")
driver.add_testcase(
"identity feature map",
hedge_models.LinearFeatureHedge(
timesteps=timesteps,
instrument_dim=tradebook.instrument_dim,
mappings=[approximators.IdentityFeatureMap] \
* (1 + (driver.cost is not None))),