def run_ica_experiment(_run, root, dataset_dict, n_components, ica_method,
fn_name):
dataset_data_dict = {}
for fold in ("train", "validation", "test"):
data = load_subjects_data(root,
subjects=dataset_dict[fold],
mode="both")
dataset_data_dict[fold] = data
datasets = dataset_data_dict
root = Path(root)
gdfs = [datasets["train"][subject]["gdf"] for subject in datasets["train"]]
joined_eeg, labels = join_gdfs_to_numpy(gdfs)
ica_transform = get_ica_transformers()[ica_method]
ica_transform.fit(joined_eeg)
results = {}
for dataset_name in datasets:
dataset_fold = datasets[dataset_name]
if not dataset_name in results:
results[dataset_name] = {}
for subject_id in dataset_fold:
gdf_data = dataset_fold[subject_id]["gdf"]
data = gdf_data.get_data().T
data_arr = ica_transform.transform(data)
start = time.time()
score = apply_pairwise_parallel(data_arr, SCORING_FN_DICT[fn_name])
duration = time.time() - start
results[dataset_name][subject_id] = {
"score": score,
"duration": duration
}
with open("./results.json", "w") as jfile:
json.dump(results, jfile, indent=4)
_run.add_artifact("results.json", content_type="json")
def run_ica_experiment(_run, method_idx):
train_epochs = BCI_IV_Comp_Dataset.load_dataset(train_filepaths,
as_epochs=True,
concatenate=True,
drop_bad=True,
return_metadata=False,
tmin=-1.,
tmax=3.)
preprocess_epochs(train_epochs, inplace=True)
test_epochs_list, test_metadata_list = BCI_IV_Comp_Dataset.load_dataset(
test_filepaths,
as_epochs=True,
concatenate=False,
drop_bad=True,
return_metadata=True,
tmin=-1.,
tmax=3.)
all_methods = get_all_methods()
methods = all_methods if method_idx is None else [all_methods[method_idx]]
name = "" if method_idx is None else "_{}".format(all_methods[method_idx])
print("Using methods", methods)
print("Loaded test files:", len(test_epochs_list))
results = dict()
for method in methods:
print("Running for method", method)
results[method] = list()
clf = make_pipeline(
AverageFrequencyBand(fmin=0.1, fmax=30, picks=PSD_PICKS),
StandardScaler(), SVC(C=4))
start = time.time()
ICA = get_ica_instance(method, n_components=ICA_N_COMPONENTS)
ICA.fit(train_epochs)
print("\tICA fitted!")
duration = time.time() - start
transformed_train_epochs = ICA.get_sources(train_epochs)
X_train, Y_train = transformed_train_epochs, transformed_train_epochs.events[:,
2]
clf.fit(X_train, Y_train)
for i, (epochs,
mdata) in enumerate(zip(test_epochs_list, test_metadata_list)):
print("\t", i, mdata["id"])
epochs = preprocess_epochs(epochs, inplace=False)
transformed_test_epochs = ICA.get_sources(epochs)
scores = dict()
signal = np.hstack(transformed_test_epochs.get_data())
for fn_name in SCORING_FN_DICT:
score = apply_pairwise_parallel(signal,
SCORING_FN_DICT[fn_name])
scores[fn_name] = score
X_test, Y_test = transformed_test_epochs, transformed_test_epochs.events[:,
2]
pred = clf.predict(X_test)
bas = balanced_accuracy_score(Y_test, pred)
results[method].append({
"id": mdata["id"],
"score": scores,
"bas": bas,
"duration": duration
})
results_filepath = f"./results{name}.json"
with open(results_filepath, "w") as json_file:
json.dump(results, json_file, indent=4)
_run.add_artifact(results_filepath, content_type="json")
def run_ica_experiment(_run, method_idx):
# filepaths = Path(r"C:\Users\paull\Documents\GIT\BCI_MsC\notebooks\BCI_Comp_IV_2a\BCICIV_2a_gdf/").glob("*T.gdf")
dataset, metadata = BCI_IV_Comp_Dataset.load_dataset(filepaths,
as_epochs=True,
concatenate=False,
drop_bad=True,
return_metadata=True,
tmin=-1.,
tmax=3.)
all_methods = get_all_methods()
methods = all_methods if method_idx is None else [all_methods[method_idx]]
name = "" if method_idx is None else "_{}".format(all_methods[method_idx])
print("Using methods", methods)
results = dict()
for method in methods:
print("Running for method", method)
clf = make_pipeline(
CSP(n_components=CSP_N_COMPONENTS), Vectorizer(), MinMaxScaler(),
LogisticRegression(penalty='l2', multi_class='auto'))
results[method] = list()
for i, (epochs, mdata) in enumerate(zip(dataset, metadata)):
print("\t", i, mdata["id"])
ICA = get_ica_instance(method, n_components=ICA_N_COMPONENTS)
start = time.time()
epochs = epochs.copy().load_data().filter(l_freq=None,
h_freq=40).resample(90.)
transformed_epochs = ICA.fit(epochs).get_sources(epochs)
duration = time.time() - start
scores = dict()
signal = np.hstack(transformed_epochs.get_data())
for fn_name in SCORING_FN_DICT:
score = apply_pairwise_parallel(signal,
SCORING_FN_DICT[fn_name])
scores[fn_name] = score
X, Y = transformed_epochs.get_data(), transformed_epochs.events[:,
2]
del epochs, transformed_epochs
try:
clf.fit(X, Y)
except Exception:
print("\t\tFailed during fit")
results[method].append({
"id": mdata["id"],
"score": None,
"bas": None,
"duration": duration
})
continue
pred = clf.predict(X)
bas = balanced_accuracy_score(Y, pred)
results[method].append({
"id": mdata["id"],
"score": scores,
"bas": bas,
"duration": duration
})
results_filepath = f"./results{name}.json"
with open(results_filepath, "w") as json_file:
json.dump(results, json_file, indent=4)
_run.add_artifact(results_filepath, content_type="json")
def run_ica_experiment(_run, method_idx):
# filepaths = Path(r"C:\Users\paull\Documents\GIT\BCI_MsC\notebooks\BCI_Comp_IV_2a\BCICIV_2a_gdf/").glob("*T.gdf")
dataset, metadata = BCI_IV_Comp_Dataset.load_dataset(filepaths,
as_epochs=True,
concatenate=False,
drop_bad=True,
return_metadata=True,
tmin=-1.,
tmax=3.)
all_methods = get_all_methods()
methods = all_methods if method_idx is None else [all_methods[method_idx]]
name = "" if method_idx is None else "_{}".format(all_methods[method_idx])
print("Using methods", methods)
results = dict()
for method in methods:
print("Running for method", method)
clf = make_pipeline(PSD(fmin=8, fmax=30, picks=PSD_PICKS), Averager(),
StandardScaler(), SVC(C=4))
results[method] = list()
for i, (epochs, mdata) in enumerate(zip(dataset, metadata)):
print("\t", i, mdata["id"])
ICA = get_ica_instance(method,
n_components=ICA_N_COMPONENTS,
random_state=SEED,
max_iter=5000)
start = time.time()
epochs = preprocess_epochs(epochs)
n_epochs = len(epochs)
epochs_idx = list(range(n_epochs))
train_epochs_idx, test_epochs_idx = train_test_split(
epochs_idx, test_size=TEST_PERCENTAGE, random_state=SEED)
train_epochs = epochs.copy().drop(test_epochs_idx)
test_epochs = epochs.copy().drop(train_epochs_idx)
ICA.fit(train_epochs)
transformed_train_epochs = ICA.get_sources(train_epochs)
transformed_test_epochs = ICA.get_sources(test_epochs)
duration = time.time() - start
scores = dict()
signal = np.hstack(transformed_test_epochs.get_data())
for fn_name in SCORING_FN_DICT:
score = apply_pairwise_parallel(signal,
SCORING_FN_DICT[fn_name])
scores[fn_name] = score
X, Y = transformed_train_epochs, transformed_train_epochs.events[:,
2]
X_test, Y_test = transformed_test_epochs, transformed_test_epochs.events[:,
2]
try:
clf.fit(X, Y)
except Exception as e:
print("\t\tFailed during fit:", str(e))
results[method].append({
"id": mdata["id"],
"score": None,
"bas": None,
"duration": duration
})
continue
pred = clf.predict(X_test)
bas = balanced_accuracy_score(Y_test, pred)
results[method].append({
"id": mdata["id"],
"score": scores,
"bas": bas,
"duration": duration
})
del epochs, transformed_train_epochs, transformed_test_epochs
results_filepath = f"./results{name}.json"
with open(results_filepath, "w") as json_file:
json.dump(results, json_file, indent=4)
_run.add_artifact(results_filepath, content_type="json")
def run_ica_experiment(_run, method_idx):
train_epochs = BCI_IV_Comp_Dataset.load_dataset(
train_filepaths,
as_epochs=True,
concatenate=True,
drop_bad=True,
return_metadata=False,
tmin=-1.,
tmax=3.
)
train_epochs.load_data().filter(l_freq=None, h_freq=40)
test_epochs_list, test_metadata_list = BCI_IV_Comp_Dataset.load_dataset(
test_filepaths,
as_epochs=True,
concatenate=False,
drop_bad=True,
return_metadata=True,
tmin=-1.,
tmax=3.
)
all_methods = get_all_methods()
methods = all_methods if method_idx is None else [all_methods[method_idx]]
name = "" if method_idx is None else "_{}".format(all_methods[method_idx])
print("Using methods", methods)
print("Loaded test files:", len(test_epochs_list))
results = dict()
for method in methods:
print("Running for method", method)
results[method] = list()
clf = make_pipeline(
CSP(n_components=CSP_N_COMPONENTS),
Vectorizer(),
MinMaxScaler(),
LogisticRegression(
penalty='l2',
multi_class='auto'
)
)
start = time.time()
ICA = get_ica_instance(method, n_components=ICA_N_COMPONENTS)
ICA.fit(train_epochs)
duration = time.time() - start
transformed_train_epochs = ICA.get_sources(train_epochs)
for i, (epochs, mdata) in enumerate(zip(test_epochs_list, test_metadata_list)):
epochs = epochs.copy().load_data()
transformed_test_epochs = ICA.get_sources(epochs)
scores = dict()
signal = np.hstack(transformed_test_epochs.get_data())
for fn_name in SCORING_FN_DICT:
score = apply_pairwise_parallel(signal, SCORING_FN_DICT[fn_name])
scores[fn_name] = score
X_train, Y_train = transformed_train_epochs.get_data(), transformed_train_epochs.events[:, 2]
X_test, Y_test = transformed_test_epochs.get_data(), transformed_test_epochs.events[:, 2]
try:
clf.fit(X_train, Y_train)
except Exception:
print("\t\tFailed during fit")
results[method].append(
{
"id": mdata["id"],
"score": None,
"bas": None,
"duration": duration
}
)
continue
pred = clf.predict(X_test)
bas = balanced_accuracy_score(Y_test, pred)
results[method].append(
{
"id": mdata["id"],
"score": scores,
"bas": bas,
"duration": duration
}
)
results_filepath = f"./results{name}.json"
with open(results_filepath, "w") as json_file:
json.dump(results, json_file, indent=4)
_run.add_artifact(results_filepath, content_type="json")