快捷導(dǎo)航

tensorflow實現(xiàn)邏輯回歸模型

更新時間：2018年09月08日 09:42:36 作者：Missayaa

這篇文章主要為大家詳細(xì)介紹了tensorflow實現(xiàn)邏輯回歸模型的相關(guān)資料，具有一定的參考價值，感興趣的小伙伴們可以參考一下

邏輯回歸模型

邏輯回歸是應(yīng)用非常廣泛的一個分類機器學(xué)習(xí)算法，它將數(shù)據(jù)擬合到一個logit函數(shù)(或者叫做logistic函數(shù))中，從而能夠完成對事件發(fā)生的概率進行預(yù)測。

import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
from tensorflow.examples.tutorials.mnist import input_data
#下載好的mnist數(shù)據(jù)集存在F:/mnist/data/中
mnist = input_data.read_data_sets('F:/mnist/data/',one_hot = True)
print(mnist.train.num_examples)
print(mnist.test.num_examples)

trainimg = mnist.train.images
trainlabel = mnist.train.labels
testimg = mnist.test.images
testlabel = mnist.test.labels

print(type(trainimg))
print(trainimg.shape,)
print(trainlabel.shape,)
print(testimg.shape,)
print(testlabel.shape,)

nsample = 5
randidx = np.random.randint(trainimg.shape[0],size = nsample)

for i in randidx:
  curr_img = np.reshape(trainimg[i,:],(28,28))
  curr_label = np.argmax(trainlabel[i,:])
  plt.matshow(curr_img,cmap=plt.get_cmap('gray'))
  plt.title(""+str(i)+"th Training Data"+"label is"+str(curr_label))
  print(""+str(i)+"th Training Data"+"label is"+str(curr_label))
  plt.show()


x = tf.placeholder("float",[None,784])
y = tf.placeholder("float",[None,10])
W = tf.Variable(tf.zeros([784,10]))
b = tf.Variable(tf.zeros([10]))

#
actv = tf.nn.softmax(tf.matmul(x,W)+b)
#計算損失
cost = tf.reduce_mean(-tf.reduce_sum(y*tf.log(actv),reduction_indices=1))
#學(xué)習(xí)率
learning_rate = 0.01
#隨機梯度下降
optm = tf.train.GradientDescentOptimizer(learning_rate).minimize(cost)

#求1位置索引值 對比預(yù)測值索引與label索引是否一樣，一樣返回True
pred = tf.equal(tf.argmax(actv,1),tf.argmax(y,1))
#tf.cast把True和false轉(zhuǎn)換為float類型 0，1
#把所有預(yù)測結(jié)果加在一起求精度
accr = tf.reduce_mean(tf.cast(pred,"float"))
init = tf.global_variables_initializer()
"""
#測試代碼 
sess = tf.InteractiveSession()
arr = np.array([[31,23,4,24,27,34],[18,3,25,4,5,6],[4,3,2,1,5,67]])
#返回數(shù)組的維數(shù) 2
print(tf.rank(arr).eval())
#返回數(shù)組的行列數(shù) [3 6]
print(tf.shape(arr).eval())
#返回數(shù)組中每一列中最大元素的索引[0 0 1 0 0 2]
print(tf.argmax(arr,0).eval())
#返回數(shù)組中每一行中最大元素的索引[5 2 5]
print(tf.argmax(arr,1).eval()) 
J"""
#把所有樣本迭代50次
training_epochs = 50
#每次迭代選擇多少樣本
batch_size = 100
display_step = 5

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

#循環(huán)迭代
for epoch in range(training_epochs):
  avg_cost = 0
  num_batch = int(mnist.train.num_examples/batch_size)
  for i in range(num_batch):
    batch_xs,batch_ys = mnist.train.next_batch(batch_size)
    sess.run(optm,feed_dict = {x:batch_xs,y:batch_ys})
    feeds = {x:batch_xs,y:batch_ys}
    avg_cost += sess.run(cost,feed_dict = feeds)/num_batch

  if epoch % display_step ==0:
    feeds_train = {x:batch_xs,y:batch_ys}
    feeds_test = {x:mnist.test.images,y:mnist.test.labels}
    train_acc = sess.run(accr,feed_dict = feeds_train)
    test_acc = sess.run(accr,feed_dict = feeds_test)
    #每五個epoch打印一次信息
    print("Epoch:%03d/%03d cost:%.9f train_acc:%.3f test_acc: %.3f" %(epoch,training_epochs,avg_cost,train_acc,test_acc))

print("Done")

程序訓(xùn)練結(jié)果如下：

Epoch:000/050 cost:1.177228655 train_acc:0.800 test_acc: 0.855
Epoch:005/050 cost:0.440933891 train_acc:0.890 test_acc: 0.894
Epoch:010/050 cost:0.383387268 train_acc:0.930 test_acc: 0.905
Epoch:015/050 cost:0.357281335 train_acc:0.930 test_acc: 0.909
Epoch:020/050 cost:0.341473956 train_acc:0.890 test_acc: 0.913
Epoch:025/050 cost:0.330586549 train_acc:0.920 test_acc: 0.915
Epoch:030/050 cost:0.322370980 train_acc:0.870 test_acc: 0.916
Epoch:035/050 cost:0.315942993 train_acc:0.940 test_acc: 0.916
Epoch:040/050 cost:0.310728854 train_acc:0.890 test_acc: 0.917
Epoch:045/050 cost:0.306357428 train_acc:0.870 test_acc: 0.918
Done

以上就是本文的全部內(nèi)容，希望對大家的學(xué)習(xí)有所幫助，也希望大家多多支持腳本之家。

您可能感興趣的文章: