def load_data_new(main_path,
main_file,
train_samples_id,
dimension,
step=2,
train=False,
preprocess=False):
ids, paths, names, sampling_rates, labels, explanations, partitions, intervals = load_file(
main_path, main_file)
train_x = np.empty((0, dimension))
train_y = np.empty((0, 2))
for i in train_samples_id:
sample_x, sample_y = load_sample(main_path + paths[i],
names[i],
labels[i],
sampling_rates[i],
explanations[i],
intervals[i],
dimension=dimension,
step=step,
train=train,
preprocess=preprocess)
train_x = np.append(train_x, sample_x, axis=0)
train_y = np.append(train_y, sample_y)
#print('%d label = %s, #sample = %d %d' %(i,labels[i], sample_x.shape[0], sample_y.shape[0]))
train_x = np.reshape(
train_x,
[train_x.shape[0], dimension, 1]) ####Hey hey check this line!
train_y = np.reshape(train_y, [train_y.shape[0], 1])
print('Total load result: #sample0=%d, #sample1=%d' %
(train_x.shape[0], train_y.shape[0]))
return train_x, train_y
def load_data(main_path,
main_file,
train_samples_id,
dimension1,
dimension2,
step=2,
train=False,
preprocess=False):
global_counter_label0 = 0
global_counter_label1 = 0
global_counter_sample0 = 0
global_counter_sample1 = 0
ids, paths, names, sampling_rates, labels, explanations, partitions, intervals = load_file(
main_path, main_file)
train_x = np.empty((0, dimension1, dimension2))
train_y = np.empty((0, 2))
for i in train_samples_id:
sample_x, sample_y, counter_label0, counter_label1, counter_sample0, counter_sample1 = load_sample_rbased(
main_path + paths[i],
names[i],
labels[i],
sampling_rates[i],
explanations[i],
dimension1=dimension1,
dimension2=dimension2,
step=step,
train=train)
train_x = np.append(train_x, sample_x, axis=0)
train_y = np.append(train_y, sample_y)
print('%d #label0 = %d, #label1=%d, #sample0=%d, #sample1=%d' %
(i, counter_label0, counter_label1, counter_sample0,
counter_sample1))
global_counter_label0 += counter_label0
global_counter_label1 += counter_label1
global_counter_sample0 += counter_sample0
global_counter_sample1 += counter_sample1
train_x = np.reshape(train_x, [train_x.shape[0], dimension2, dimension1
]) ####Hey hey check this line!
print('%#label0 = %d, #label1=%d, #sample0=%d, #sample1=%d' %
(global_counter_label0, global_counter_label1,
global_counter_sample0, global_counter_sample1))
return train_x, train_y
71, 173, 331, 327, 336, 159, 35, 178, 38, 66, 240, 18, 261, 371, 200,
238, 372, 126, 135, 256, 242, 107, 258, 32, 152, 153, 277, 286, 1, 232,
165, 223, 202, 333, 123, 300, 99, 53, 292, 63, 282, 212, 376, 183, 6,
12, 309, 56, 175, 14, 194, 218, 145, 297, 325, 114, 278, 298, 230, 79,
206, 203, 125, 7
]))
'''
main_path = '/Users/Zeynab/'
#main_file = 'My data/In use/Data_v960213.csv'
main_file = 'My data/In use/Data_v960412.csv'
'''
main_path = '/home/mll/Golgooni/'
main_file = 'My data/In use/Data_v960412.csv'
ids, paths, names, sampling_rates, labels, explanations, partitions, intervals = load_file(
main_path, main_file)
############################# Set parameters #################################
raw_dimension = 4000
mother_wavelet = 'db4'
wav_level = 8
num_coefficient = 4
wav_dimension = 0
#wav_dimension = 22 + 22 + 38 + 69
rnn_layer = 'LSTM'
rnn_hidden_node = 3
rnn_dropout = 0.4
batch_size = 1
epochs = 20