def simu_fit_test(arma_params,
ll,
order,
n_train_samples,
n_test_samples,
data='arma',
embedding='lead_lag'):
inputSig = InputSig(data, embedding, order, ll=ll, arma_params=arma_params)
train_X, train_y = get_input_X_y(inputSig,
n_train_samples,
start_row,
n_processes=n_processes)
max_train_X = np.amax(np.absolute(train_X), axis=0)
train_X = train_X / max_train_X
test_X, test_y = get_input_X_y(inputSig,
n_test_samples,
start_row + n_train_samples,
n_processes=n_processes)
test_X = test_X / max_train_X
# Fit and evaluate algorithm
learnSig = LearnSig(algo, inputSig, n_processes=n_processes)
learnSig.train(train_X, train_y)
test_results = learnSig.evaluate(test_X, test_y, metrics=['error_l2'])
return (test_results['error_l2'])
n_processes=n_processes)
valid_X = valid_X / max_train_X
test_X = test_X / max_train_X
# Fit and evaluate algorithm
learnSig = LearnSig(algo, inputSig, n_processes=n_processes)
# Train with custom parameters, for example for Urban Sound best random forest
#custom_params={'n_estimators':460,'max_depth':30,'max_features':500}
#learnSig.train(
# train_X,train_y,valid_X=valid_X,valid_y=valid_y,params=custom_params)
learnSig.train(train_X, train_y, valid_X=valid_X, valid_y=valid_y)
test_results = learnSig.evaluate(test_X,
test_y,
metrics=['accuracy', 'f1_score'])
print(test_results)
# Write the results in a separate text file.
results_path = os.path.join(results_dir, learnSig.model_name + ".txt")
file = open(results_path, "w")
file.write("Dataset: %s \n" % (data))
file.write("Algorithm: %s \n" % (algo))
file.write("Embedding: %s \n" % (embedding))
file.write("Signature order: %i \n" % (order))
file.write("Number of training samples: %i \n" % (n_train_samples))
file.write("Number of validation samples: %i \n" % (n_valid_samples))
file.write("Number of test samples: %i \n" % (n_test_samples))
n_processes=n_processes)
valid_X = valid_X / max_train_X
test_X, test_y = get_input_X_y(inputSig,
n_test_samples,
start_row + n_train_samples +
n_valid_samples,
n_processes=n_processes)
valid_X = valid_X / max_train_X
test_X = test_X / max_train_X
for algo in algo_list:
# Fit and evaluate algorithm
learnSig = LearnSig(algo, inputSig, n_processes=n_processes)
learnSig.train(train_X, train_y, valid_X=valid_X, valid_y=valid_y)
test_results = learnSig.evaluate(test_X, test_y, metrics=metrics)
print(test_results)
results_df = results_df.append(
{
'embedding': embedding,
'algo': algo,
'order': order,
'n_features': train_X.shape[1]
},
ignore_index=True)
print(results_df)
for metric in metrics:
results_df[metric].iloc[-1] = test_results[metric]
# Write the results in a separate text file.