def store(output_dir, cv_method):
''' INPUT
output_dir: is the directory where to store the keymetrics (ranking, weights and accuracies dataframes)
cv_method : 5_fold , 10_fold or LOO
OUTPUT
stored pickle dataframes in the given directory
'''
# Initialize empty dataframes
pool_FS = [
reliefF
] #,lap_score,ll_l21,ls_l21,UDFS,fisher_score,chi_square,gini_index,SPEC]
labels = [
'reliefF'
] #,'lap_score','ll_l21','ls_l21','UDFS','fisher_score','chi_square','gini_index','SPEC']#,'Boratapy']
dataframe_ranking = pd.DataFrame(index=num_fea, columns=labels)
dataframe_weights = pd.DataFrame(index=num_fea, columns=labels)
dataframe_accuracies = pd.DataFrame(index=num_fea, columns=labels)
#matrix_=np.zeros((50,589*3))
for i in range(len(pool_FS)):
for k in num_fea:
ranking__, acc__, weight__ = training(cv_method, k, pool_FS[i], X,
y)
#ranking__,acc__=training(kf5,k,pool_FS[i],X,y)
#ranking__,acc__,=training(kf5,k,pool_FS[i])
dataframe_ranking[labels[i]][k] = ranking__
dataframe_weights[labels[i]][k] = weight__
dataframe_accuracies[labels[i]][k] = acc__
#dataframe_ranking_5fold=dataframe_ranking.copy()
#dataframe_weights_5fold=dataframe_weights.copy()
#dataframe_accuracies_5fold=dataframe_accuracies.copy()
name1, name2, name3 = name_dataframe(X, cv_method)
dataframe_accuracies.to_pickle(output_dir + name1)
dataframe_ranking.to_pickle(output_dir + name2)
dataframe_weights.to_pickle(output_dir + name3)
# You should have received a copy of the GNU General Public License
# along with Foobar. If not, see <http://www.gnu.org/licenses/>.
#
# (c) Junya Kaneko <*****@*****.**>
from matplotlib import pyplot as plt
from nn.networks import Classifier
from dataset import MnistTrainingDataset, MnistTestDataset
from helpers import training, test, draw_W_histories, draw_mean_se_history, draw_cpr_history
if __name__ == '__main__':
# Load MNIST dataset
training_dataset = MnistTrainingDataset('./mnist', 1, 0)
test_dataset = MnistTestDataset('./mnist', 1, 0)
# Create Deep Neural Network
classifier = Classifier('rectifier', training_dataset.img_size, 'se', 0.15)
classifier.add_layer('rectifier', 200)
classifier.add_layer('rectifier', 10)
W_histories, mean_se_history, cpr_history = training(classifier, training_dataset, 400)
draw_W_histories(W_histories, classifier.name, training_dataset.name)
draw_mean_se_history(mean_se_history, classifier.name, training_dataset.name)
draw_cpr_history(cpr_history, classifier.name, training_dataset.name)
print(test(classifier, test_dataset))
plt.show()