def fitness(data, label, zero_index, answer):
out = np.zeros(5)
if label == 3:
out[0] = 1
elif label == 5:
out[1] = 1
elif label == 6:
out[2] = 1
elif label == 7:
out[3] = 1
elif label == 9:
out[4] = 1
dd = data
for e in range(20):
dd[zero_index + e] = answer[e]
label_arr = out
model = load_model('./model.h5')
g = [dd]
res = model.predict(funcs.extract(g))
label_classifier = res[0] #
score = 0
for f in range(len(label_arr)):
d = label_arr[f] - label_classifier[f]
if d < 0:
d *= -1
score += d
return 5 - score
def extract():
if "url" not in request.values:
return response(400, msg="Missing parameter.")
url = request.values["url"]
return funcs.extract(url)
for item in label_train:
out = np.zeros(5)
if item == 3:
out[0] = 1
elif item == 5:
out[1] = 1
elif item == 6:
out[2] = 1
elif item == 7:
out[3] = 1
elif item == 9:
out[4] = 1
output_data.append(out)
output_data = np.asarray(output_data)
data_features = funcs.extract(data_train)
print("===========================")
print('Start NN')
model = Sequential()
model.add(Dense(100, activation='relu'))
model.add(Dense(200, activation='relu'))
model.add(Dense(4000, activation='sigmoid'))
model.add(Dense(100, activation='softmax'))
model.add(Dense(5, activation='sigmoid'))
model.compile(loss='mean_squared_error',
optimizer='sgd',
metrics=['accuracy'])
print("Start fit data")
trainedModel = model.fit(data_features,
output_data,
epochs=100,
output_data = []
for item in t_label:
out = np.zeros(5)
if item == 3:
out[0] = 1
elif item == 5:
out[1] = 1
elif item == 6:
out[2] = 1
elif item == 7:
out[3] = 1
elif item == 9:
out[4] = 1
output_data.append(out)
output_data = np.asarray(output_data)
f = funcs.extract(t_data)
res = model.evaluate(f, output_data)
print(res)
print(model.metrics_names)
res = model.predict(f)
for i, item in enumerate(res):
idxMax = np.argmax(item)
predict = 0
if idxMax == 0:
predict = 3
elif idxMax == 1:
predict = 5
elif idxMax == 2:
import numpy as np
from keras.models import load_model
import funcs
model = load_model('./model.h5')
print("get test data")
data_test, label_test = funcs.get_test_data()
f = funcs.extract(data_test)
output_data = []
for item in label_test:
out = np.zeros(5)
if item == 3:
out[0] = 1
elif item == 5:
out[1] = 1
elif item == 6:
out[2] = 1
elif item == 7:
out[3] = 1
elif item == 9:
out[4] = 1
output_data.append(out)
output_data = np.asarray(output_data)
res = model.evaluate(f, output_data)
print(res)
print(model.metrics_names)
res = model.predict(f)