tensorflow实现线性回归和逻辑回归

关于线性回归和逻辑回归的原理和python实现，请左转：逻辑回归、线性回归。

这里就直接贴代码了。

线性回归：

# -*- coding: utf-8 -*-
"""
Created on Thu Aug 30 09:40:50 2018

@author: 96jie
"""

import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
#数据
num_points = 1000
vectors_set = []
for i in range(num_points):
    x1 = np.random.normal(0.0, 0.55)
    y1 = x1 * 0.1 + 0.3 + np.random.normal(0.0, 0.03)
    vectors_set.append([x1, y1])
x = [v[0] for v in vectors_set]
y = [v[1] for v in vectors_set]
#随机生成theta
theta = tf.Variable(tf.random_uniform([2],-1,1),name='theta')
#计算损失
y11 = theta[0]*x+ theta[1]
loss = tf.reduce_mean(tf.square(y11 - y), name='loss')
#梯度下降
optimizer = tf.train.GradientDescentOptimizer(0.01)
#训练得到最小梯度
train = optimizer.minimize(loss, name='train')

sess = tf.Session()
init = tf.global_variables_initializer()
sess.run(init)
#迭代
for step in range(1000):
    sess.run(train)
print(sess.run(theta))

plt.figure(figsize=(35,35))
plt.scatter(x,y,marker='o')
plt.plot(x,sess.run(y11))
plt.xticks(fontsize=40)
plt.yticks(fontsize=40)
plt.xlabel('特征X',fontsize=40)
plt.ylabel('Y',fontsize=40)
plt.title('结果',fontsize=40)
plt.savefig("tf结果.jpg")

输出结果：

逻辑回归：

# -*- coding: utf-8 -*-
"""
Created on Thu Aug 30 11:27:55 2018

@author: asus
"""
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt 

label = []
feature = np.zeros([32561,123])
f = open(r'D:\python_test\test\test\样本\train.txt')
line = f.readline()
a = 0
while line:
    data = []
    for i in line.split( ):
        data.append(i);
    for i in data[1:]:
        j = i.split(":")
        feature[a][int(j[0]) - 1] = int(j[1])
    if data[0] in '+1':
       label.append(1) 
    else:
       label.append(0)         
    line = f.readline()
    a += 1
f.close
n = len(label)
label = np.mat(label)


#构建训练集和测试集
label1 = label[:,20001:32561]
label = label[:,0:20000]
feature1 = feature[20001:32561]
feature = feature[0:20000]
#feature1 = np.insert(feature1, 0, values=one1, axis=1)
#feature = np.insert(feature, 0, values=one, axis=1)

#参数
numclass = 1
Iter = 6000
inputSize = 123

#指定好x和y的大小
X = tf.placeholder(tf.float32, shape = [None, inputSize])
y = tf.placeholder(tf.float32, shape = [None, numclass])
#参数初始化
W1 = tf.Variable(tf.ones([123, 1]))
B1 = tf.Variable(tf.constant(0.1), [numclass])

#损失函数
y_pred = 1 / (1 + tf.exp(-tf.matmul(X, W1) + B1))
loss = tf.reduce_mean(- y * tf.log(y_pred) - (1 - y) * tf.log(1 - y_pred))
#训练
opt = tf.train.GradientDescentOptimizer(learning_rate = 0.05)
train = opt.minimize(loss)
y1 = tf.round(y_pred)
#计算准确率
correct_prediction = tf.equal(y1, y)
accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))

init = tf.global_variables_initializer()
sess = tf.Session()
sess.run(init)

for i in range(Iter):
    batchInput = feature
    batchLabels = label.transpose()
    sess.run(train, feed_dict={X: batchInput, y: batchLabels})
    if i%1000 == 0:
        train_accuracy = accuracy.eval(session=sess, feed_dict={X: batchInput, y: batchLabels})
        print ("step %d, training accuracy %g"%(i, train_accuracy))
        
#测试集 
testAccuracy = sess.run(accuracy, feed_dict={X: feature1, y: label1.transpose()})
print ("test accuracy %g"%(testAccuracy))

结果：

step 0, training accuracy 0.23805
step 1000, training accuracy 0.82775
step 2000, training accuracy 0.8402
step 3000, training accuracy 0.84385
step 4000, training accuracy 0.84585
step 5000, training accuracy 0.84585
test accuracy 0.844984

逻辑回归的数据用的依旧是之前的数据。

代码数据