tensorflow 实现 svr

import tensorflow as tf

import numpy as np

import matplotlib.pyplot as plt

rng = np.random

 

 

# Parameters

learning_rate = 0.01

training_epochs = 1000

display_step = 50

batch_size = 10

 

# Training Data

train_X = np.asarray([3.3,4.4,5.5,6.71,6.93,4.168,9.779,6.182,7.59,2.167,

                         7.042,10.791,5.313,7.997,5.654,9.27,3.1])

train_Y = np.asarray([1.7,2.76,2.09,3.19,1.694,1.573,3.366,2.596,2.53,1.221,

                         2.827,3.465,1.65,2.904,2.42,2.94,1.3])

n_samples = train_X.shape[0]

 

# tf Graph Input

X = tf.placeholder(tf.float32)

Y = tf.placeholder(tf.float32)

A = tf.Variable(tf.constant(1))

 

# Set model weights

W = tf.Variable(rng.randn(), name="weight")

b = tf.Variable(rng.randn(), name="bias")

e =  tf.constant([0.0])

 

 

# Construct a linear model

pred = tf.add(tf.multiply(W, X), b)

 

 

# Mean squared error

cost = tf.reduce_mean(tf.maximum(0., tf.subtract(tf.abs(tf.subtract(pred, Y)), e)))

# Gradient descent

optimizer = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost)

# Initialize the variables (i.e. assign their default value)

init = tf.global_variables_initializer()

 

# Start training

with tf.Session() as sess:

    sess.run(init)

 

    # Fit all training data

    for epoch in range(training_epochs):

        rand_index = np.random.choice(len(train_X), size=batch_size)

        rand_x = train_X[rand_index]

        rand_y = train_Y[rand_index]

        sess.run(optimizer, feed_dict={X: rand_x[:,None], Y: rand_y[:,None]})

 

        #Display logs per epoch step

        if (epoch+1) % display_step == 0:

            c = sess.run(cost, feed_dict={X: train_X[:,None], Y:train_Y[:,None]})

            print "Epoch:", '%04d' % (epoch+1), "cost=", "{:.9f}".format(c), \

                "W=", sess.run(W), "b=", sess.run(b)

 

    print "Optimization Finished!"

    training_cost = sess.run(cost, feed_dict={X: train_X[:,None], Y: train_Y[:,None]})

    print "Training cost=", training_cost, "W=", sess.run(W), "b=", sess.run(b), '\n'

 

    #Graphic display

    plt.plot(train_X, train_Y, 'ro', label='Original data')

    plt.plot(train_X, sess.run(W) * train_X + sess.run(b), label='Fitted line')

    plt.plot(train_X, sess.run(W) * train_X + sess.run(b) - sess.run(e), "--", label='down line')

    plt.plot(train_X, sess.run(W) * train_X + sess.run(b) + sess.run(e), "--", label='up line')

    plt.legend()

    plt.show()