-
Notifications
You must be signed in to change notification settings - Fork 2
/
testing_SPM_gm+wm_SVM_1_per_site.py
91 lines (68 loc) · 3.41 KB
/
testing_SPM_gm+wm_SVM_1_per_site.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
"""
Experiment using Linear SVM on WM+GM SPM data.
Results:
MAE: Mean(SD) = 4.95453854
"""
from pathlib import Path
import pandas as pd
import numpy as np
from sklearn.model_selection import KFold, GridSearchCV
from sklearn.svm import SVR
PROJECT_ROOT = Path.cwd()
# --------------------------------------------------------------------------
random_seed = 42
np.random.seed(random_seed)
# --------------------------------------------------------------------------
# Create experiment's output directory
testing_dir = PROJECT_ROOT / 'output' / 'testing'
testing_dir.mkdir(exist_ok=True)
# --------------------------------------------------------------------------
# Input data directory (plz, feel free to use NAN shared folder)
gram_matrix_path = PROJECT_ROOT / 'data' / 'gram' / 'wm+gm_testing.csv'
train_demographic_path = PROJECT_ROOT / 'data' / 'PAC2019_BrainAge_Training.csv'
test_demographic_path = PROJECT_ROOT / 'data' / 'PAC2019_BrainAge_Test_Upload.csv'
# Reading data. If necessary, create new reader in helper_functions.
train_demographic_df = pd.read_csv(train_demographic_path)
test_demographic_df = pd.read_csv(test_demographic_path)
gram_df = pd.read_csv(gram_matrix_path, index_col='subject_ID')
x_train = gram_df[list(train_demographic_df['subject_ID'])]
x_train = x_train.reindex(list(train_demographic_df['subject_ID']))
x_train = x_train.values
x_test = gram_df[list(train_demographic_df['subject_ID'])]
x_test = x_test.reindex(list(test_demographic_df['subject_ID']))
x_test = x_test.values
# --------------------------------------------------------------------------
# Using only age
y_train = train_demographic_df['age'].values
sites_train = train_demographic_df['site'].values
sites_test = test_demographic_df['site'].values
for i_site in np.unique(sites_train):
x_train_red = x_train[sites_train==i_site, :][:,sites_train==i_site]
y_train_red = y_train[sites_train==i_site]
x_test_red = x_test[sites_test==i_site, :][:,sites_train==i_site]
testing_name = 'SPM_wm+gm_SVM_1_per_site_{:}'.format(i_site)
# --------------------------------------------------------------------------
# Model
clf = SVR(kernel='precomputed')
# --------------------------------------------------------------------------
# Model selection
# Search space
param_grid = {'C': [2 ** -6, 2 ** -5, 2 ** -4, 2 ** -3, 2 ** -2, 2 ** -1, 2 ** 0, 2 ** 1]}
# Gridsearch
internal_cv = KFold(n_splits=5)
grid_cv = GridSearchCV(estimator=clf,
param_grid=param_grid,
cv=internal_cv,
scoring='neg_mean_absolute_error',
verbose=1)
# --------------------------------------------------------------------------
grid_result = grid_cv.fit(x_train_red, y_train_red)
# --------------------------------------------------------------------------
best_regressor = grid_cv.best_estimator_
# --------------------------------------------------------------------------
y_test_predicted = best_regressor.predict(x_test_red)
# --------------------------------------------------------------------------
predictions_df = pd.DataFrame(test_demographic_df.loc[test_demographic_df['site']==i_site]['subject_ID'])
predictions_df['age'] = y_test_predicted
predictions_df.to_csv(testing_dir / (testing_name + '.csv'), index=False)
np.save(testing_dir / (testing_name + '.npy'), grid_result.best_score_)