return delay_cat
if __name__ == '__main__':
model_path = '../model/signal_embeded_cat.h5'
coll_prefix = Read_Collection_train()
fdata_train = GetData(coll_prefix)
ProcMLData = mongo2pd(fdata_train,
time_step=15,
special_list=['SS_Subval'])
MLdf = pd.DataFrame(ProcMLData)
import delay_analyzier as delay_a
delay_a.delay_analysis(MLdf)
y_train_mean, y_train_logmean, y_train_max, y_train_logmax = label_generator(
MLdf)
y_train_logmean = pd.get_dummies(
transform_delay2category(pd.DataFrame(y_train_mean)))
model = k.Sequential()
model.add(Dense(64, input_dim=112, activation='selu'))
model.add(Dense(32, activation='selu'))
model.add(Dense(16, activation='selu'))
model.add(Dense(8, activation='selu'))
model.add(Dense(4, activation='softmax'))
#model = k.models.load_model(model_path)
if __name__ == '__main__':
coll_prefix = Read_Collection_train()
fdata_train = GetData(coll_prefix)
import mongo2pd_v3 as mpd
ProcMLData = mpd.mongo2pd(fdata_train, time_step=15)
coll_prefix = Read_Collection_test()
fdata_test = GetData(coll_prefix)
ProcMLData_test = mpd.mongo2pd(fdata_test, time_step=15)
MLdf = pd.DataFrame(ProcMLData)
MLdf_test = pd.DataFrame(ProcMLData_test)
import delay_analyzier as delay_a
delay_a.delay_analysis(MLdf)
delay_a.delay_analysis(MLdf_test)
'''
Prepare training set and validation set
'''
y_train_mean, y_train_logmean, y_train_max, y_train_logmax = label_generator(
MLdf)
y_test_mean, y_test_logmean, y_test_max, y_test_logmax = label_generator(
MLdf_test)
train = MLdf
valid = MLdf_test
y_train = y_train_logmean
y_valid = y_test_logmean
#y_train = y_train_mean