def datasets_table():
files = experiments_helper.find_crossvalidation_datasets()
valid_files = [
f for f in files
if experiments_helper.validate_dataset(f, [2, 3, 5, 10]) is not None
]
for idx_file, file in enumerate(valid_files):
dataset, result_folder = experiments_helper.validate_dataset(
file, [2, 3, 5, 10])
basefolder = os.path.basename(os.path.dirname(file))
print(basefolder + " " + str(dataset.shape[0]) + "x" +
str(dataset.shape[1]))
def test_crossvalidation():
files = experiments_helper.find_crossvalidation_datasets()
folds = experiments_helper.TRAINING_FOLDS
patience = experiments_helper.PATIENCE
string_file = "Dataset," + ','.join(["GBN_Validation_"+ str(f) + "_" + str(p) for f in folds for p in patience]) +\
",BIC,BGe"
print(string_file)
for file in files:
x = experiments_helper.validate_dataset(
file, experiments_helper.TRAINING_FOLDS)
if x is None:
continue
else:
dataset, result_folder = x
validation_gaussian = np.full(
(experiments_helper.EVALUATION_FOLDS, len(folds), len(patience)),
np.nan)
bic_gaussian = np.full((experiments_helper.EVALUATION_FOLDS, ), np.nan)
bge_gaussian = np.full((experiments_helper.EVALUATION_FOLDS, ), np.nan)
for (idx_fold, (train_indices, test_indices)) in enumerate(
KFold(experiments_helper.EVALUATION_FOLDS,
shuffle=True,
random_state=experiments_helper.SEED).split(dataset)):
train_dataset = dataset.iloc[train_indices, :]
test_dataset = dataset.iloc[test_indices, :]
validation_gaussian[idx_fold] = test_validation_gaussian(
train_dataset, test_dataset, folds, patience, result_folder,
idx_fold)
bic_gaussian[idx_fold] = test_bic_gaussian(train_dataset,
test_dataset,
result_folder, idx_fold)
bge_gaussian[idx_fold] = test_bge_gaussian(train_dataset,
test_dataset,
result_folder, idx_fold)
sum_validation_gaussian = validation_gaussian.sum(axis=0)
sum_bic = bic_gaussian.sum(axis=0)
sum_bge = bge_gaussian.sum(axis=0)
basefolder = os.path.basename(os.path.dirname(file))
new_line = basefolder
for idx_f, f in enumerate(folds):
for idx_p, p in enumerate(patience):
new_line += "," + str(sum_validation_gaussian[idx_f, idx_p])
new_line += "," + str(sum_bic)
new_line += "," + str(sum_bge)
print(new_line)
string_file += '\n' + new_line
with open('results_hc_gbn.csv', 'w') as f:
f.write(string_file)
def train_crossvalidation_file(file, folds, patience):
x = experiments_helper.validate_dataset(file, folds)
if x is None:
return
else:
dataset, result_folder = x
if not os.path.exists(result_folder):
os.mkdir(result_folder)
print(file)
with mp.Pool(processes=experiments_helper.EVALUATION_FOLDS) as p:
p.starmap(run_validation_gaussian, [(dataset.iloc[train_indices,:], folds, patience, result_folder, idx_fold)
for (idx_fold, (train_indices, test_indices)) in
enumerate(KFold(experiments_helper.EVALUATION_FOLDS, shuffle=True,
random_state=experiments_helper.SEED).split(dataset))]
)
with mp.Pool(processes=experiments_helper.EVALUATION_FOLDS) as p:
p.starmap(run_bic_gaussian, [(dataset.iloc[train_indices,:], result_folder, idx_fold)
for (idx_fold, (train_indices, test_indices)) in
enumerate(KFold(experiments_helper.EVALUATION_FOLDS, shuffle=True,
random_state=0).split(dataset))]
)
with mp.Pool(processes=experiments_helper.EVALUATION_FOLDS) as p:
p.starmap(run_bge_gaussian, [(dataset.iloc[train_indices,:], result_folder, idx_fold)
for (idx_fold, (train_indices, test_indices)) in
enumerate(KFold(experiments_helper.EVALUATION_FOLDS,
shuffle=True, random_state=0).split(dataset))]
)
def print_datasets():
datasets = experiments_helper.find_crossvalidation_datasets()
for dataset in datasets:
basefolder = os.path.basename(os.path.dirname(dataset))
x = experiments_helper.validate_dataset(
dataset, experiments_helper.TRAINING_FOLDS)
if x is None:
continue
else:
dataset, _ = x
print("{} & {} & {}\\\\".format(basefolder, dataset.shape[0],
dataset.shape[1]))
def num_instances_df():
files = experiments_helper.find_crossvalidation_datasets()
names = []
num_instances = []
for file in files:
x = experiments_helper.validate_dataset(
file, experiments_helper.TRAINING_FOLDS)
if x is not None:
dataset, result_folder = x
names.append(os.path.basename(result_folder))
num_instances.append(dataset.shape[0])
return pd.DataFrame({"Dataset": names, "N": num_instances})
def test_crossvalidation():
files = experiments_helper.find_crossvalidation_datasets()
folds = experiments_helper.TRAINING_FOLDS
patience = experiments_helper.PATIENCE
string_file = "Dataset," + ','.join(["SPBN_CKDE_PC_LC_" + str(f) + "_" + str(p) for f in folds for p in patience]) +\
"," + ','.join(["SPBN_CKDE_PC_RCOT_" + str(f) + "_" + str(p) for f in folds for p in patience])
print(string_file)
for file in files:
x = experiments_helper.validate_dataset(file, experiments_helper.TRAINING_FOLDS)
if x is None:
continue
else:
dataset, result_folder = x
spbn_lc_score = np.full((experiments_helper.EVALUATION_FOLDS, len(folds), len(patience)), np.nan)
spbn_rcot_score = np.full((experiments_helper.EVALUATION_FOLDS, len(folds), len(patience)), np.nan)
for (idx_fold, (train_indices, test_indices)) in enumerate(KFold(experiments_helper.EVALUATION_FOLDS, shuffle=True,
random_state=experiments_helper.SEED).split(dataset)):
train_dataset = dataset.iloc[train_indices,:]
test_dataset = dataset.iloc[test_indices,:]
spbn_lc_score[idx_fold] = test_pc_lc_spbn(train_dataset, test_dataset, folds, patience,
result_folder, idx_fold)
spbn_rcot_score[idx_fold] = test_pc_rcot_spbn(train_dataset, test_dataset, folds, patience,
result_folder, idx_fold)
sum_spbn_lc_score = spbn_lc_score.sum(axis=0)
sum_spbn_rcot_score = spbn_rcot_score.sum(axis=0)
basefolder = os.path.basename(os.path.dirname(file))
new_line = basefolder
for idx_f, f in enumerate(folds):
for idx_p, p in enumerate(patience):
new_line += "," + str(sum_spbn_lc_score[idx_f, idx_p])
for idx_f, f in enumerate(folds):
for idx_p, p in enumerate(patience):
new_line += "," + str(sum_spbn_rcot_score[idx_f, idx_p])
print(new_line)
string_file += '\n' + new_line
with open('results_pc_spbn_ckde.csv', 'w') as f:
f.write(string_file)
def test_crossvalidation():
files = experiments_helper.find_crossvalidation_datasets()
string_file = "Dataset,GBN_PC_LC,GBN_PC_RCOT"
print(string_file)
for file in files:
x = experiments_helper.validate_dataset(
file, experiments_helper.TRAINING_FOLDS)
if x is None:
continue
else:
dataset, result_folder = x
gbn_lc_score = np.full((experiments_helper.EVALUATION_FOLDS, ), np.nan)
gbn_rcot_score = np.full((experiments_helper.EVALUATION_FOLDS, ),
np.nan)
for (idx_fold, (train_indices, test_indices)) in enumerate(
KFold(experiments_helper.EVALUATION_FOLDS,
shuffle=True,
random_state=experiments_helper.SEED).split(dataset)):
train_dataset = dataset.iloc[train_indices, :]
test_dataset = dataset.iloc[test_indices, :]
gbn_lc_score[idx_fold] = test_pc_lc_gbn(train_dataset,
test_dataset,
result_folder, idx_fold)
gbn_rcot_score[idx_fold] = test_pc_rcot_gbn(
train_dataset, test_dataset, result_folder, idx_fold)
sum_gbn_lc_score = gbn_lc_score.sum(axis=0)
sum_gbn_rcot_score = gbn_rcot_score.sum(axis=0)
basefolder = os.path.basename(os.path.dirname(file))
new_line = basefolder + "," + str(sum_gbn_lc_score) + "," + str(
sum_gbn_rcot_score)
print(new_line)
string_file += '\n' + new_line
with open('results_pc_gbn.csv', 'w') as f:
f.write(string_file)
def kdeness_ckde():
folds = experiments_helper.TRAINING_FOLDS
patience = experiments_helper.PATIENCE
files = experiments_helper.find_crossvalidation_datasets()
valid_files = [
f for f in files
if experiments_helper.validate_dataset(f, folds) is not None
]
n_ckde = np.full((len(valid_files), len(folds), 3, 10), np.nan)
datasets = []
n_vars = []
for idx_file, file in enumerate(valid_files):
x = experiments_helper.validate_dataset(
file, experiments_helper.TRAINING_FOLDS)
dataset, result_folder = x
basefolder = os.path.basename(os.path.dirname(file))
datasets.append(basefolder)
n_vars.append(dataset.shape[1])
for idx_f, f in enumerate(experiments_helper.TRAINING_FOLDS):
for idx_fold in range(10):
models_folder = result_folder + '/HillClimbing/SPBN_CKDE/Validation_' + str(
f) + '_5/' + str(idx_fold)
all_models = sorted(glob.glob(models_folder + '/*.pickle'))
final_model = load(all_models[-1])
n_ckde[idx_file, idx_f, 0, idx_fold] = \
sum(map(lambda kv: kv[1] == NodeType.CKDE, final_model.node_types().items()))
for idx_f, f in enumerate(experiments_helper.TRAINING_FOLDS):
for idx_fold in range(10):
models_folder = result_folder + '/PC/SPBN_CKDE/LinearCorrelation/Validation_' + str(
f) + '_5/' + str(idx_fold)
all_models = sorted(glob.glob(models_folder + '/*.pickle'))
final_model = load(all_models[-1])
n_ckde[idx_file, idx_f, 1, idx_fold] = \
sum(map(lambda kv: kv[1] == NodeType.CKDE, final_model.node_types().items()))
for idx_f, f in enumerate(experiments_helper.TRAINING_FOLDS):
for idx_fold in range(10):
models_folder = result_folder + '/PC/SPBN_CKDE/RCoT/Validation_' + str(
f) + '_5/' + str(idx_fold)
all_models = sorted(glob.glob(models_folder + '/*.pickle'))
final_model = load(all_models[-1])
n_ckde[idx_file, idx_f, 2, idx_fold] = \
sum(map(lambda kv: kv[1] == NodeType.CKDE, final_model.node_types().items()))
mean_ckde = np.mean(n_ckde, axis=3).reshape(len(valid_files), -1)
algorithms = ["HC", "PC-PLC", "PC-RCoT"]
names = [
"CKDE_" + str(f) + "_" + algorithm for f in folds
for algorithm in algorithms
]
df = pd.DataFrame(mean_ckde, columns=names, index=datasets)
df['n_vars'] = n_vars
for f in folds:
for algorithm in algorithms:
df['%CKDE_' + str(f) + "_" +
algorithm] = df.loc[:, 'CKDE_' + str(f) + "_" +
algorithm] / df.loc[:, 'n_vars']
N = df.shape[0]
ind = np.arange(N)
num_bars = len(folds) * len(algorithms)
width = (1 - 0.3) / num_bars
fig = plt.figure()
ax = fig.add_subplot(111)
offset = 0
b = []
color = {
algorithms[0]: "#729CF5",
algorithms[1]: "#FFB346",
algorithms[2]: "#B5EA7F"
}
for f in folds:
for algorithm in algorithms:
t = ax.bar(ind + width * offset,
df['%CKDE_' + str(f) + "_" + algorithm].to_numpy(),
width,
align='edge',
linewidth=0.5,
edgecolor="black",
color=color[algorithm])
offset += 1
b.append(t)
ax.set_ylabel('Ratio of CKDE variables')
ax.set_xticks(ind + (1 - 0.3) / 2)
ax.set_xticklabels(df.index)
ax.tick_params(axis='x', rotation=90)
plt.legend([t[0] for t in b], algorithms)
tikzplotlib.save("plots/kdeness.tex",
standalone=True,
axis_width="25cm",
axis_height="10cm")