def prediction(time_step=20):
X = tf.placeholder(tf.float32, shape=[None, time_step, input_size])
mean, std, test_x, test_y = get_test_data(time_step)
with tf.variable_scope("sec_lstm", reuse=tf.AUTO_REUSE):
pred, _ = lstm(X)
saver = tf.train.Saver(tf.global_variables())
with tf.Session() as sess:
# 参数恢复
module_file = tf.train.latest_checkpoint('model_save2')
saver.restore(sess, module_file)
test_predict = []
for step in range(len(test_x) - 1):
prob = sess.run(pred, feed_dict={X: [test_x[step]], keep_prob: 1})
predict = prob.reshape((-1))
test_predict.extend(predict)
test_y = np.array(test_y) * std[7] + mean[7]
test_predict = np.array(test_predict) * std[7] + mean[7]
acc = np.average(
np.abs(test_predict - test_y[:len(test_predict)]) /
test_y[:len(test_predict)]) # 偏差程度
print("The accuracy of this predict:", acc)
# 以折线图表示结果
plt.figure()
plt.plot(
list(range(len(test_predict))),
test_predict,
color='b',
)
plt.plot(list(range(len(test_y))), test_y, color='r')
plt.show()
def plot_results(xlog, ulog):
t = [0.25 * x for x in range(xlog.shape[1])]
plt.figure()
plt.subplot(311)
plt.plot(t, xlog[3, :], label="altitude", color="b")
plt.ylabel('altitude')
plt.subplot(312)
plt.plot(t, xlog[1, :], label="pitch angle", color="r")
plt.ylabel('pitch angle')
plt.subplot(313)
plt.plot(t, ulog[0, :], label="elevator angle", color="g")
plt.ylabel('elevator angle')
plt.show()
training_epochs = 1000
# 损失函数: subtract张量的减法
loss = tf.reduce_sum(
tf.maximum(0., tf.subtract(tf.abs(tf.subtract(curr_y, y)), epsilon)))
optimizer = tf.train.GradientDescentOptimizer(learning_rate)
train = optimizer.minimize(loss)
sess = tf.Session()
sess.run(tf.global_variables_initializer())
for i in range(training_epochs):
sess.run(train, {x: t_x, y: t_y})
if i % 20 == 0:
print(i, sess.run([a, b, loss], {x: t_x, y: t_y}))
a_val = sess.run(a)
b_val = sess.run(b)
print("this model is y=", a_val, "* x +", b_val)
# sess.close()
y_learned = t_x * a_val + b_val
plt.plot(t_x, t_y, "k.")
plt.plot(t_x, y_learned, "g-")
linewidth = sess.run(epsilon)
plt.plot(t_x, y_learned + linewidth, "r--")
plt.plot(t_x, y_learned - linewidth, "r--")
plt.savefig("svm.png")
plt.show()
