def run_optimization_tests(algorithm, input_files, key_file, is_weighted, max_eval):
inputs = prepare_tw_input(input_files, is_weighted)
search_space = get_search_space_by_algorithm(algorithm)
LOGGER.info(
"algorithm:{}, key_file: {}, weighted= {}, # of files: {}, "
"max_eval={}".format(algorithm, key_file, is_weighted,
len(input_files), max_eval))
optimize(objective, inputs, key_file, search_space, max_eval)
def run_kmeans(paths: List[Dict[str, str]], params):
message = 'Running K-Means experiments'
print(message + '...')
logging.info(message)
results_to_save = '### K-Means experiments results\n'
results_to_save += 'Except K, the other parameters are the default ones (eg. euclidean distance)\n'
for path in paths:
results_to_save += f'{path["name"]} dataset\n'
X = pd.read_csv(os.path.join('datasets', path['X']))
Y = pd.read_csv(os.path.join('datasets', path['Y']), header=None)
# Optimization of K
alg_params = {
'name': path['name'],
'vis_dims': 2,
'fig_save_path': params.output_path
}
results = optimize(X=X.values,
algorithm=KMeans,
algorithm_params=alg_params,
metric='calinski_harabasz_score',
metric_params={'X': X.values},
k_values=list(range(2, 15)),
goal='minimize')
results_to_save = generate_results(X, Y, results, results_to_save)
with open(os.path.join(params.output_path, 'results.md'), 'a') as f:
f.write(results_to_save)