5 用神经网络解决非线性问题

最新推荐文章于 2022-11-01 15:09:20 发布

大春SSC

最新推荐文章于 2022-11-01 15:09:20 发布

阅读量924

点赞数

分类专栏：深度学习基于Python语言和TensorFlow平台

深度学习基于Python语言和TensorFlow平台专栏收录该内容

14 篇文章

订阅专栏

博客介绍了使用随机数据进行训练的相关内容，包括score2a程序及结果，还提到加入偏移量b以加快训练过程，有score2b的程序和结果，最后涉及批量生成随机训练数据，给出了score2c的程序和结果。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

使用随机数据进行训练

5.1 score2a
程序：

import tensorflow as tf
import random

random.seed()

x = tf.placeholder(dtype=tf.float32)
yTrain = tf.placeholder(dtype=tf.float32)

w = tf.Variable(tf.zeros([3]), dtype=tf.float32)

wn = tf.nn.softmax(w)

n1 = wn * x

n2 = tf.reduce_sum(n1)

y = tf.nn.sigmoid(n2)

loss = tf.abs(yTrain - y)

optimizer = tf.train.RMSPropOptimizer(0.1)

train = optimizer.minimize(loss)

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

for i in range(5):
    xData = [int(random.random() * 8 + 93), int(random.random() * 8 + 93), int(random.random() * 8 + 93)]
    xAll = xData[0] * 0.6 + xData[1] * 0.3 + xData[2] * 0.1

    if xAll >= 95:
        yTrainData = 1
    else:
        yTrainData = 0

    result = sess.run([train, x, yTrain, w, n2, y, loss], feed_dict={x: xData, yTrain: yTrainData})
    print(result)

    xData = [int(random.random() * 41 + 60), int(random.random() * 41 + 60), int(random.random() * 41 + 60)]
    xAll = xData[0] * 0.6 + xData[1] * 0.3 + xData[2] * 0.1

    if xAll >= 95:
        yTrainData = 1
    else:
        yTrainData = 0

    result = sess.run([train, x, yTrain, w, n2, y, loss], feed_dict={x: xData, yTrain: yTrainData})
    print(result)

结果：

[None, array([94., 93., 96.], dtype=float32), array(0., dtype=float32), array([0., 0., 0.], dtype=float32), 94.333336, 1.0, 1.0]
[None, array([93., 60., 99.], dtype=float32), array(0., dtype=float32), array([0., 0., 0.], dtype=float32), 84.0, 1.0, 1.0]
[None, array([ 94., 100.,  98.], dtype=float32), array(1., dtype=float32), array([0., 0., 0.], dtype=float32), 97.33334, 1.0, 0.0]
[None, array([68., 73., 83.], dtype=float32), array(0., dtype=float32), array([0., 0., 0.], dtype=float32), 74.66667, 1.0, 1.0]
[None, array([96., 95., 94.], dtype=float32), array(1., dtype=float32), array([0., 0., 0.], dtype=float32), 95.0, 1.0, 0.0]
[None, array([86., 94., 99.], dtype=float32), array(0., dtype=float32), array([0., 0., 0.], dtype=float32), 93.0, 1.0, 1.0]
[None, array([ 93., 100.,  97.], dtype=float32), array(1., dtype=float32), array([0., 0., 0.], dtype=float32), 96.66667, 1.0, 0.0]
[None, array([91., 81., 96.], dtype=float32), array(0., dtype=float32), array([0., 0., 0.], dtype=float32), 89.333336, 1.0, 1.0]
[None, array([96., 97., 99.], dtype=float32), array(1., dtype=float32), array([0., 0., 0.], dtype=float32), 97.333336, 1.0, 0.0]
[None, array([92., 90., 88.], dtype=float32), array(0., dtype=float32), array([0., 0., 0.], dtype=float32), 90.0, 1.0, 1.0]

加入偏移量b加快训练过程

5.2 score2b
程序：

import tensorflow as tf
import random

random.seed()

x = tf.placeholder(dtype=tf.float32)
yTrain = tf.placeholder(dtype=tf.float32)

w = tf.Variable(tf.zeros([3]), dtype=tf.float32)
b = tf.Variable(80, dtype=tf.float32)

wn = tf.nn.softmax(w)

n1 = wn * x

n2 = tf.reduce_sum(n1) - b

y = tf.nn.sigmoid(n2)

loss = tf.abs(yTrain - y)

optimizer = tf.train.RMSPropOptimizer(0.1)

train = optimizer.minimize(loss)

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

for i in range(500):
    xData = [int(random.random() * 8 + 93), int(random.random() * 8 + 93), int(random.random() * 8 + 93)]
    xAll = xData[0] * 0.6 + xData[1] * 0.3 + xData[2] * 0.1

    if xAll >= 95:
        yTrainData = 1
    else:
        yTrainData = 0

    result = sess.run([train, x, yTrain, wn, b, n2, y, loss], feed_dict={x: xData, yTrain: yTrainData})
    print(result)

    xData = [int(random.random() * 41 + 60), int(random.random() * 41 + 60), int(random.random() * 41 + 60)]
    xAll = xData[0] * 0.6 + xData[1] * 0.3 + xData[2] * 0.1

    if xAll >= 95:
        yTrainData = 1
    else:
        yTrainData = 0

    result = sess.run([train, x, yTrain, wn, b, n2, y, loss], feed_dict={x: xData, yTrain: yTrainData})
    print(result)

结果：

...
[None, array([94., 62., 86.], dtype=float32), array(0., dtype=float32), array([0.6192946 , 0.25602525, 0.12468012], dtype=float32), 93.66678, -8.857033, 0.00014235664, 0.00014235664]
[None, array([ 99.,  94., 100.], dtype=float32), array(1., dtype=float32), array([0.6185921 , 0.25666544, 0.12474249], dtype=float32), 93.666954, 4.1744614, 0.9848496, 0.015150428]
[None, array([65., 61., 67.], dtype=float32), array(0., dtype=float32), array([0.628845  , 0.24586184, 0.12529323], dtype=float32), 93.647736, -29.380592, 0.0, 0.0]
[None, array([100.,  97.,  93.], dtype=float32), array(1., dtype=float32), array([0.628845  , 0.24586184, 0.12529323], dtype=float32), 93.647736, 4.737625, 0.9913167, 0.008683324]
[None, array([84., 92., 81.], dtype=float32), array(0., dtype=float32), array([0.6364523 , 0.24166633, 0.12188137], dtype=float32), 93.63542, -8.067734, 0.00031339467, 0.00031339467]

批量生成随机训练数据

5.3 score2c
程序：

import tensorflow as tf
import random
import numpy as np

random.seed()

rowCount = 5

xData = np.full(shape=(rowCount, 3), fill_value=0, dtype=np.float32)
yTrainData = np.full(shape=rowCount, fill_value=0, dtype=np.float32)

goodCount = 0

# 生成随机训练数据的循环
for i in range(rowCount):
    xData[i] = [int(random.random() * 11 + 90), int(random.random() * 11 + 90), int(random.random() * 11 + 90)]
    xAll = xData[i][0] * 0.6 + xData[i][1] * 0.3 + xData[i][2] * 0.1

    if xAll >= 95:
        yTrainData[i] = 1
        goodCount = goodCount + 1
    else:
        yTrainData[i] = 0

print("xData=%s" % xData)
print("yTrainData=%s" % yTrainData)
print("goodCount=%d" % goodCount)

x = tf.placeholder(dtype=tf.float32)
yTrain = tf.placeholder(dtype=tf.float32)

w = tf.Variable(tf.zeros([3]), dtype=tf.float32)
b = tf.Variable(80, dtype=tf.float32)

wn = tf.nn.softmax(w)

n1 = wn * x

n2 = tf.reduce_sum(n1) - b

y = tf.nn.sigmoid(n2)

loss = tf.abs(yTrain - y)

optimizer = tf.train.RMSPropOptimizer(0.1)

train = optimizer.minimize(loss)

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

for i in range(2):
    for j in range(rowCount):
        result = sess.run([train, x, yTrain, wn, b, n2, y, loss], feed_dict={x: xData[j], yTrain: yTrainData[j]})
        print(result)

结果：

xData=[[94. 93. 97.]
 [96. 92. 91.]
 [94. 95. 98.]
 [96. 97. 93.]
 [91. 97. 91.]]
yTrainData=[0. 0. 0. 1. 0.]
goodCount=1

[None, array([94., 93., 97.], dtype=float32), array(0., dtype=float32), array([0.33333334, 0.33333334, 0.33333334], dtype=float32), 80.0, 14.666672, 0.9999995, 0.9999995]
[None, array([96., 92., 91.], dtype=float32), array(0., dtype=float32), array([0.33333334, 0.33333334, 0.33333334], dtype=float32), 80.0, 13.0, 0.99999774, 0.99999774]
[None, array([94., 95., 98.], dtype=float32), array(0., dtype=float32), array([0.33333325, 0.33333337, 0.33333337], dtype=float32), 80.0, 15.666672, 0.9999999, 0.9999999]
[None, array([96., 97., 93.], dtype=float32), array(1., dtype=float32), array([0.33333325, 0.33333337, 0.33333337], dtype=float32), 80.0, 15.333328, 0.99999976, 2.3841858e-07]
[None, array([91., 97., 91.], dtype=float32), array(0., dtype=float32), array([0.33333325, 0.33333337, 0.33333337], dtype=float32), 80.0, 13.0, 0.99999774, 0.99999774]
[None, array([94., 93., 97.], dtype=float32), array(0., dtype=float32), array([0.33333334, 0.33333325, 0.33333343], dtype=float32), 80.0, 14.666672, 0.9999995, 0.9999995]
[None, array([96., 92., 91.], dtype=float32), array(0., dtype=float32), array([0.3333333 , 0.33333325, 0.3333334 ], dtype=float32), 80.0, 13.0, 0.99999774, 0.99999774]
[None, array([94., 95., 98.], dtype=float32), array(0., dtype=float32), array([0.33333322, 0.3333333 , 0.3333335 ], dtype=float32), 80.0, 15.666672, 0.9999999, 0.9999999]
[None, array([96., 97., 93.], dtype=float32), array(1., dtype=float32), array([0.33333322, 0.3333333 , 0.3333335 ], dtype=float32), 80.0, 15.333336, 0.99999976, 2.3841858e-07]
[None, array([91., 97., 91.], dtype=float32), array(0., dtype=float32), array([0.33333322, 0.33333328, 0.33333346], dtype=float32), 80.0, 12.999992, 0.99999774, 0.99999774]