def meta_fitting(data_to_predict_local, data_input_2, strategy_dictionary):
fitting_inputs, fitting_targets = input_processing(data_to_predict_local, data_input_2, strategy_dictionary)
error = []
train_indices, test_indices = train_test_indices(fitting_inputs, strategy_dictionary['train_test_ratio'])
if strategy_dictionary['ml_mode'] == 'svm':
fitting_dictionary, error = svm_fitting(
fitting_inputs, fitting_targets, train_indices, test_indices, strategy_dictionary)
elif strategy_dictionary['ml_mode'] == 'randomforest':
fitting_dictionary, error = random_forest_fitting(
fitting_inputs, fitting_targets, train_indices, test_indices, strategy_dictionary)
elif strategy_dictionary['ml_mode'] == 'adaboost':
fitting_dictionary, error = adaboost_fitting(
fitting_inputs, fitting_targets, train_indices, test_indices, strategy_dictionary)
elif strategy_dictionary['ml_mode'] == 'gradientboosting':
fitting_dictionary, error = gradient_boosting_fitting(
fitting_inputs, fitting_targets, train_indices, test_indices, strategy_dictionary)
elif strategy_dictionary['ml_mode'] == 'extratreesfitting':
fitting_dictionary, error = extra_trees_fitting(
fitting_inputs, fitting_targets, train_indices, test_indices, strategy_dictionary)
fitting_dictionary['train_indices'] = train_indices
fitting_dictionary['test_indices'] = test_indices
fitting_dictionary['error'] = error
return fitting_dictionary
def fit_tensorflow(strategy_dictionary):
toc = tic()
data_to_predict, data_2 = import_data(strategy_dictionary)
fitting_inputs, fitting_targets = input_processing(data_to_predict, data_2,
strategy_dictionary)
train_indices, test_indices = train_test_indices(
fitting_inputs, strategy_dictionary['train_test_ratio'])
if strategy_dictionary['sequence_flag']:
fitting_dictionary, error = tensorflow_sequence_fitting(
'/home/thomas/test', train_indices, test_indices, fitting_inputs,
fitting_targets, strategy_dictionary)
else:
fitting_dictionary, error = tensorflow_fitting(train_indices,
test_indices,
fitting_inputs,
fitting_targets)
fitting_dictionary['train_indices'] = train_indices
fitting_dictionary['test_indices'] = test_indices
fitting_dictionary = post_process_training_results(strategy_dictionary,
fitting_dictionary,
data_to_predict)
profit_factor = output_strategy_results(strategy_dictionary,
fitting_dictionary,
data_to_predict, toc)
return fitting_dictionary, data_to_predict, error, profit_factor