def evaluate(self, afile):
f2b = File2Buffer()
f2b.file_load(afile)
index_of_max_value = tf.argmax(self.hypothesis, 1)
hit_record = tf.equal(index_of_max_value, tf.argmax(self.Y_one_hot, 1))
recognition_rate = tf.reduce_mean(tf.cast(hit_record, tf.float32))
#
acc = self.sess.run(recognition_rate,
feed_dict={
self.X: f2b.x_data,
self.Y: f2b.y_data
})
print("Acc: {:.2%}".format(acc * 100))
def evaluate_file_one_hot(self, afile, num_of_class):
f2b = File2Buffer()
f2b.file_load(afile)
Y_one_hot_reshaped = self.Y_2_one_hot(f2b.y_data, num_of_class)
index = tf.argmax(self.hypothesis,
1) # op for returning index of a max value
correct_prediction = tf.equal(index, tf.argmax(Y_one_hot_reshaped, 1))
accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))
acc = self.sess.run(accuracy,
feed_dict={
self.X: f2b.x_data,
self.Y: f2b.y_data
})
print("Acc: {:.2%}".format(acc))
def learn_with_file(self, afile, total_loop, check_step):
f2b = File2Buffer()
f2b.file_load(afile)
#print(f2b.x_data, f2b.y_data)
self.learn(f2b.x_data, f2b.y_data, total_loop, check_step)
def load_file(self, afile):
f2b = File2Buffer()
f2b.file_load(afile)
return f2b.x_data, f2b.y_data
def learn_from_file(self, afile, total_loop, check_step):
f2b = File2Buffer()
f2b.file_load(afile)
#f2b.print_info()
self.learn(f2b.x_data, f2b.y_data, total_loop, check_step)
output = tf.sigmoid(output)
self.set_hypothesis(output)
self.set_cost_function(NNType.LOGISTIC)
self.set_optimizer(NNType.GRADIENT_DESCENT, l_rate=0.01)
def my_log(self, i, x_data, y_data):
pass
gildong = MVLogisticRegression4Diabetes()
gildong.learn_with_file('data-03-diabetes.csv', 10000, 200) #10000, 200
gildong.test_sigmoid(
[[0.176471, 0.155779, 0, 0, 0, 0.052161, -0.952178, -0.733333]])
f2b = File2Buffer()
f2b.file_load('data-03-diabetes.csv')
gildong.evaluate_sigmoid(f2b.x_data, f2b.y_data)
gildong.show_error()
'''
1200 0.654603
1400 0.640737
1600 0.62813
1800 0.616668
2000 0.606246
[[ 0.6939525]]
[ 0.55056906]
[ 0.71810943]
[ 0.72589421]
[ 0.58412576]