TF之NN之回归预测：利用NN算法(RelU)实现根据三个自变量预测一个因变量的回归问题

版权声明：本文为博主原创文章，未经博主允许不得转载。 https://blog.csdn.net/qq_41185868/article/details/84559637

TF之NN：基于TF利用NN算法实现根据三个自变量预测一个因变量的回归问题

TF之NN之回归预测：利用NN算法(RelU)实现基于30行样本(每个样本含有18列参数包括label)预测一个新样本值

实验数据

说明：利用前四年的数据建立回归模型，并对第五年进行预测。

设计思路

 

输出结果

[code]loss is: 913.6623
loss is: 781206160000.0
loss is: 9006693000.0
loss is: 103840136.0
loss is: 1197209.2
loss is: 13816.644
loss is: 173.0564
loss is: 15.756571
loss is: 13.9430275
loss is: 13.922119
loss is: 13.921878
loss is: 13.921875
loss is: 13.921875
loss is: 13.921875
loss is: 13.921875
input is:[120, 5, 85]
output is:15.375002

[code]y_list is [[65.0], [64.0], [64.0], [61.0], [59.0], [61.0], [62.0], [60.0], [61.0], [61.0], [59.0], [61.0], [61.0], [60.0], [60.0], [60.0], [59.0], [61.0], [60.0], [59.0], [60.0], [60.0], [59.0], [60.0], [60.0], [60.0], [64.0], [61.0], [59.0], [59.0]]
x_pred is [[6.0, 30.0, 0.5, 3.12, 3.13, 364.0, 452.0, 473.0, 1858.0, 1996.0, 2036.0, 0.23, 0.47, 0.5, 0.0, 146.0, 149.0]]

loss is: 747890.9
loss is: 6.603946e+26
loss is: 7.613832e+24
loss is: 8.778148e+22
loss is: 1.0120521e+21
loss is: 1.1668169e+19
loss is: 1.3452473e+17
loss is: 1550963400000000.0
loss is: 17881387000000.0
loss is: 206158180000.0
loss is: 2376841700.0
loss is: 27403098.0
loss is: 315938.72
loss is: 3645.1162
loss is: 44.61735
loss is: 3.1063914
loss is: 2.627804
loss is: 2.6222868
loss is: 2.622223
loss is: 2.6222224
input is:[6.0, 30.0, 0.5, 3.12, 3.13, 364.0, 452.0, 473.0, 1858.0, 1996.0, 2036.0, 0.23, 0.47, 0.5, 0.0, 146.0, 149.0]
output is:60.66666

 

实现代码

[code]import numpy as np
import tensorflow as tf

x = [[80,3,50],[90,8,70],[180,20,120],[140,16,90]]
y = [[11],[12.5],[20],[18]]
# y = [11,12.5,20,18]
x_pred = [[120,5,85]]

# dataset = np.loadtxt("data/20181127test04.csv", delimiter=",")
# # split into input (X) and output (Y) variables
# x = dataset[0:7,0:17]
# print(x)
# y = dataset[0:7,17]
# print(y)
# x_pred = dataset[7,0:17]
# print(x_pred)

tf_x = tf.placeholder(tf.float32, [None,3])     # input x
tf_y = tf.placeholder(tf.float32, [None,1])     # input y

print(tf_x)

# neural network layers
l1 = tf.layers.dense(tf_x, 100, tf.nn.relu)          # hidden layer  18*8
output = tf.layers.dense(l1, 1)                      # output layer

loss = tf.losses.mean_squared_error(tf_y, output)   # compute cost
optimizer = tf.train.GradientDescentOptimizer(learning_rate=0.1)
train_op = optimizer.minimize(loss)

sess = tf.Session()                                 # control training and others
sess.run(tf.global_variables_initializer())         # initialize var in graph

for step in range(150):
# train and net output
_, l, pred = sess.run([train_op, loss, output], {tf_x: x, tf_y: y})
if step % 10 == 0:
print('loss is: ' + str(l))
# print('prediction is:' + str(pred))

output_pred = sess.run(output,{tf_x:x_pred})
print('input is:' + str(x_pred[0][:]))
print('output is:' + str(output_pred[0][0]))

 

阅读更多