def experiment2(datafilepath, columns_info_file_path):
dataset = DataSet(*load_dataset(datafilepath, columns_info_file_path))
dataset_size = dataset.size
split_sizes = {}
split_sizes['testing_dataset_size'] = 75
split_sizes['validation_dataset_size'] = 30
remaining_data_size = dataset_size - split_sizes[
'testing_dataset_size'] - split_sizes['validation_dataset_size']
split_sizes['training_dataset_size'] = remaining_data_size
complete_training_dataset, testing_dataset, validation_dataset = datasetfilters.split_dataset_by_sizes(
dataset, split_sizes)
initial_ratio = 0.3
current_ratio = initial_ratio
iterations_results = []
iteration = 1
while (current_ratio < 1):
print("Iteration no:", iteration)
split_sizes['training_dataset_size'] = int(current_ratio *
remaining_data_size)
#training_dataset, testing_dataset, validation_dataset = datasetfilters.split_dataset_by_sizes(dataset, split_sizes)
training_dataset = complete_training_dataset.subset(
[i for i in range(0, split_sizes['training_dataset_size'])])
results = experiment_iteration(training_dataset, testing_dataset,
validation_dataset)
results['iteration_results']['training_dataset_size'] = split_sizes[
'training_dataset_size']
iterations_results.append(results)
iteration = iteration + 1
current_ratio = current_ratio + 0.05
print_iterations_results_table2(iterations_results)
def experiment1(datafilepath, columns_info_file_path, no_of_iterations):
dataset = DataSet(*load_dataset(datafilepath, columns_info_file_path))
ratios = [0.6, 0.35, 0.05]
iterations_results = []
for iteration in range(1, no_of_iterations + 1):
print("Iteration no:", iteration)
training_dataset, testing_dataset, validation_dataset = datasetfilters.split_dataset(
dataset, ratios)
results = experiment_iteration(training_dataset, testing_dataset,
validation_dataset)
results['iteration_index'] = iteration
iterations_results.append(results)
plot_decison_tree_iterations_results(iterations_results)
print_iterations_results_table(iterations_results)
)
#sys.exit(0)
elif 2 not in indexes:
print(
'data info file (.info) path is not present in arguments list... ending program... bye.. bye... '
)
#sys.exit(0)
else:
datafilepath = sys.argv[1]
columns_info_file_path = sys.argv[2]
print('Input data set is ', datasetname)
mode = 'run'
if (True):
dataset = DataSet(*load_dataset(datafilepath, columns_info_file_path))
dataset_size = dataset.size
testing_size = 50
training_size = dataset_size - testing_size
ratios = [training_size / dataset_size, testing_size / dataset_size, 0]
print('{:-<200}'.format('-'))
print('Dataset loaded successfully, datafile:', datafilepath,
' datainfofilepath:', columns_info_file_path)
training_stop_criteria = {'type': 'weights_change', 'value': 0.00005}
training_stop_criteria['max_iterations'] = 1000
dataset_description = {'datasetname': datasetname}
m = 10
if (mode == 'run'):
#nayes_bayes_experiment(dataset_description, dataset, ratios, m)
multi_classifier_roc_experiment(dataset_description, dataset,