神经网络_使用TensorFlow预测气温

已于 2024-09-16 15:36:12 修改
阅读量629 收藏 15
点赞数 14
分类专栏： 算法、机器学习和数学 文章标签： 神经网络 算法 人工智能 tensorflow 预测气温
于 2024-09-16 08:11:07 首次发布
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/feivirus/article/details/142297711
版权 
import datetime
import numpy as np
import pandas as pd 
import matplotlib.pyplot as plt
import tensorflow as tf
from tensorflow.keras import layers
import tensorflow.keras 
import warnings
from sklearn import preprocessing
warnings.filterwarnings("ignore")
%matplotlib inline

1.数据预处理

1.1导入数据

temperatures = pd.read_csv('tf_temperature.csv')
print("数据维度", temperatures.shape)
temperatures.head()


数据维度 (348, 9)
yearmonthdayweektemp_2temp_1averageactualfriend
0201611Fri454545.64529
1201612Sat444545.74461
2201613Sun454445.84156
3201614Mon444145.94053
4201615Tues414046.04441

数据说明

  • temp_2 前天的最高温度值
  • temp_1 昨天的最高温度值
  • averate 历史上，每年这一天的平均最高温度值
  • actual 标签值，当天的真实的最高温度值
  • friend 朋友猜测的温度值,忽略

1.2 处理时间数据

years = temperatures['year']
months = temperatures['month']
days = temperatures['day']

#转为datetime格式
temp_days = [str(int(year)) + '-' + str(int(month)) + '-' + str(int(day)) for year, month, day in zip(years, months, days)]
temp_days = [datetime.datetime.strptime(date, '%Y-%m-%d') for date in temp_days]
print(temp_days[:5])

[datetime.datetime(2016, 1, 1, 0, 0), datetime.datetime(2016, 1, 2, 0, 0), datetime.datetime(2016, 1, 3, 0, 0), datetime.datetime(2016, 1, 4, 0, 0), datetime.datetime(2016, 1, 5, 0, 0)]

1.3 原始数据画图展示

#画图风格
plt.style.use('fivethirtyeight')
#布局
fig,((plt1,plt2),(plt3,plt4)) = plt.subplots(nrows=2, ncols=2,figsize=(10,10))
fig.autofmt_xdate(rotation = 45)
#标签值
plt1.plot(temp_days, temperatures['actual'])
plt1.set_xlabel('')
plt1.set_ylabel('Temperature')
plt1.set_title('Max Temp')

plt2.plot(temp_days, temperatures['temp_1'])
plt2.set_xlabel('')
plt2.set_ylabel('Temperature')
plt2.set_title('Yesterday Max Temp')

plt3.plot(temp_days, temperatures['temp_2'])
plt3.set_xlabel('')
plt3.set_ylabel('Temperature')
plt3.set_title('The Day Before Yesterday Temp')

plt4.plot(temp_days, temperatures['friend'])
plt4.set_xlabel('')
plt4.set_ylabel('Temperature')
plt4.set_title('Friend Estimate')

Text(0.5, 1.0, 'Friend Estimate')

在这里插入图片描述

1.4 ont hot编码

# one hot 编码
temperatures = pd.get_dummies(temperatures, dtype=int)
temperatures.head(5)
yearmonthdaytemp_2temp_1averageactualfriendweek_Friweek_Monweek_Satweek_Sunweek_Thursweek_Tuesweek_Wed
0201611454545.645291000000
1201612444545.744610010000
2201613454445.841560001000
3201614444145.940530100000
4201615414046.044410000010

1.5转换格式

labels = np.array(temperatures['actual'])
#去掉标签特征
temperatures = temperatures.drop('actual', axis=1)
temperatures_columns = list(temperatures.columns)
temperatures = np.array(temperatures)
print("temperatures shape", temperatures.shape)
#标准化
input_temperatures = preprocessing.StandardScaler().fit_transform(temperatures)
print(input_temperatures[0])

temperatures shape (348, 14)
[ 0.         -1.5678393  -1.65682171 -1.48452388 -1.49443549 -1.3470703
 -1.98891668  2.44131112 -0.40482045 -0.40961596 -0.40482045 -0.40482045
 -0.41913682 -0.40482045]

2.模型实现

2.1常用模型参数说明

  • activation: 激活函数，一般用relu
  • kernal_initializer,bias_initializer: 权重与偏置参数的初始化方法。有时不收敛可以换个初始化方法
  • kernel_regularize,bias_regularize: 是否加入正则化
  • inputs: 输入，自己指定或者网络自动选
  • units: 神经元个数

2.2 模型实现

#添加模型隐层
model = tf.keras.Sequential()
model.add(layers.Dense(16))
model.add(layers.Dense(32))
model.add(layers.Dense(1))

#配置模型
model.compile(optimizer=tf.keras.optimizers.SGD(0.001),
             loss="mean_squared_error")

#训练
model.fit(input_temperatures, labels, validation_split=0.25, epochs=10, batch_size=64)

Epoch 1/10
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 30ms/step - loss: 4321.0967 - val_loss: 2869.1660
Epoch 2/10
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 2461.2610 - val_loss: 3275.8066
Epoch 3/10
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 159.0226 - val_loss: 2383.7563
Epoch 4/10
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 104.4375 - val_loss: 1969.7244
Epoch 5/10
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 563.5617 - val_loss: 1057.4402
Epoch 6/10
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 60.7969 - val_loss: 868.3708
Epoch 7/10
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 133.9573 - val_loss: 622.4547
Epoch 8/10
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 53.3111 - val_loss: 625.0286
Epoch 9/10
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 43.0903 - val_loss: 393.4365
Epoch 10/10
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 125.3164 - val_loss: 294.2254





<keras.src.callbacks.history.History at 0x24426001990>

#模型 loss损失比较大，需要优化，降低loss
model.summary()

Model: "sequential_1"

┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓
┃ Layer (type)                         ┃ Output Shape                ┃         Param # ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩
│ dense_3 (Dense)                      │ (None, 16)                  │             240 │
├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤
│ dense_4 (Dense)                      │ (None, 32)                  │             544 │
├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤
│ dense_5 (Dense)                      │ (None, 1)                   │              33 │
└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘

 Total params: 819 (3.20 KB)

 Trainable params: 817 (3.19 KB)

 Non-trainable params: 0 (0.00 B)

 Optimizer params: 2 (12.00 B)

2.3 模型优化

2.3.1 更改初始化方法，增加训练次数

#更改初始化方法,训练次数从10次到100次
model = tf.keras.Sequential()
model.add(layers.Dense(16, kernel_initializer='random_normal'))
model.add(layers.Dense(32, kernel_initializer='random_normal'))
model.add(layers.Dense(1, kernel_initializer='random_normal'))

model.compile(optimizer=tf.keras.optimizers.SGD(0.001),
             loss='mean_squared_error')
model.fit(input_temperatures, labels, validation_split=0.25, epochs=100, batch_size=64)

Epoch 1/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 27ms/step - loss: 4341.7227 - val_loss: 2871.0271
Epoch 2/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 4269.7217 - val_loss: 2792.2722
Epoch 3/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 4259.1123 - val_loss: 2687.3899
Epoch 4/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 4066.6692 - val_loss: 2498.6802
Epoch 5/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 3462.5059 - val_loss: 2466.5791
Epoch 6/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 581.6287 - val_loss: 1843.6437
Epoch 7/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 90.2633 - val_loss: 1232.4553
Epoch 8/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 79.9981 - val_loss: 881.9412
Epoch 9/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 41.1886 - val_loss: 674.9135
Epoch 10/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 42.3882 - val_loss: 481.0506
Epoch 11/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 44.9377 - val_loss: 391.7140
Epoch 12/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 38.8932 - val_loss: 378.5648
Epoch 13/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 34.8383 - val_loss: 303.6538
Epoch 14/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 38.4258 - val_loss: 220.1120
Epoch 15/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 30.6331 - val_loss: 177.5809
Epoch 16/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step - loss: 25.3459 - val_loss: 151.5260
Epoch 17/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 31.5611 - val_loss: 190.2681
Epoch 18/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step - loss: 33.8859 - val_loss: 146.4823
Epoch 19/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 31.4742 - val_loss: 118.5758
Epoch 20/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 28.0468 - val_loss: 64.5958
Epoch 21/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 28.4419 - val_loss: 163.6804
Epoch 22/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 98.4275 - val_loss: 64.5287
Epoch 23/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 75.8126 - val_loss: 137.2301
Epoch 24/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 59.2577 - val_loss: 98.7127
Epoch 25/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 3ms/step - loss: 38.4441 - val_loss: 124.6540
Epoch 26/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 71.7989 - val_loss: 36.2800
Epoch 27/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 33.7251 - val_loss: 33.0628
Epoch 28/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 37.3204 - val_loss: 42.4540
Epoch 29/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 65.5749 - val_loss: 54.9257
Epoch 30/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 52.5856 - val_loss: 37.0988
Epoch 31/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 29.4801 - val_loss: 71.2360
Epoch 32/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 108.5365 - val_loss: 35.8163
Epoch 33/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 54.8995 - val_loss: 27.9930
Epoch 34/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 24.7711 - val_loss: 42.6019
Epoch 35/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 41.1184 - val_loss: 34.6733
Epoch 36/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step - loss: 27.6766 - val_loss: 31.9773
Epoch 37/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 28.4529 - val_loss: 19.7460
Epoch 38/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 35.8389 - val_loss: 35.2612
Epoch 39/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step - loss: 28.3160 - val_loss: 42.6479
Epoch 40/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 106.6452 - val_loss: 24.0267
Epoch 41/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 23.6601 - val_loss: 19.3733
Epoch 42/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step - loss: 28.4145 - val_loss: 20.0657
Epoch 43/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step - loss: 30.2289 - val_loss: 22.5946
Epoch 44/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 31.1034 - val_loss: 17.4016
Epoch 45/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step - loss: 31.8360 - val_loss: 82.5532
Epoch 46/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 107.9492 - val_loss: 21.9323
Epoch 47/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 27.2920 - val_loss: 21.0179
Epoch 48/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 25.8970 - val_loss: 22.3299
Epoch 49/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 25.3085 - val_loss: 37.8563
Epoch 50/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 40.0604 - val_loss: 18.3307
Epoch 51/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 25.8000 - val_loss: 21.0737
Epoch 52/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 25.8438 - val_loss: 43.8438
Epoch 53/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 9ms/step - loss: 33.8743 - val_loss: 33.6241
Epoch 54/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 62.6123 - val_loss: 20.5208
Epoch 55/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 31.7541 - val_loss: 42.6262
Epoch 56/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 28.6584 - val_loss: 24.4605
Epoch 57/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 31.3697 - val_loss: 22.8914
Epoch 58/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 24.2317 - val_loss: 42.9988
Epoch 59/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 51.6604 - val_loss: 42.7188
Epoch 60/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 42.3546 - val_loss: 27.8267
Epoch 61/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step - loss: 32.0020 - val_loss: 32.2499
Epoch 62/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 36.6703 - val_loss: 18.0478
Epoch 63/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 33.0255 - val_loss: 41.6100
Epoch 64/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 52.2897 - val_loss: 64.4359
Epoch 65/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 81.0884 - val_loss: 30.0917
Epoch 66/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 27.2542 - val_loss: 20.2524
Epoch 67/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 29.1652 - val_loss: 54.8341
Epoch 68/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 67.4783 - val_loss: 22.0776


Epoch 69/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 32.1612 - val_loss: 27.9906
Epoch 70/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 9ms/step - loss: 34.0457 - val_loss: 23.1129
Epoch 71/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 24.4963 - val_loss: 26.8693
Epoch 72/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step - loss: 25.6235 - val_loss: 18.3104
Epoch 73/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 25.7223 - val_loss: 30.3030
Epoch 74/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 41.1724 - val_loss: 22.1834
Epoch 75/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 25.4155 - val_loss: 18.0785
Epoch 76/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 37.8443 - val_loss: 22.2262
Epoch 77/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 25.2023 - val_loss: 24.7404
Epoch 78/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 29.4313 - val_loss: 28.2004
Epoch 79/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 32.5010 - val_loss: 34.4478
Epoch 80/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 28.1699 - val_loss: 19.9688
Epoch 81/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 33.1135 - val_loss: 30.1685
Epoch 82/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 33.0771 - val_loss: 33.2978
Epoch 83/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 28.9822 - val_loss: 36.3891
Epoch 84/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 100.3202 - val_loss: 18.8064
Epoch 85/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 25.1813 - val_loss: 42.5619
Epoch 86/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step - loss: 50.4812 - val_loss: 33.3420
Epoch 87/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 36.7241 - val_loss: 44.3068
Epoch 88/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 37.9590 - val_loss: 23.1378
Epoch 89/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step - loss: 26.4826 - val_loss: 21.8839
Epoch 90/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 31.0949 - val_loss: 23.2860
Epoch 91/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 44.1823 - val_loss: 33.2644
Epoch 92/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 36.3967 - val_loss: 26.2339
Epoch 93/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 32.2672 - val_loss: 27.6349
Epoch 94/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 26.5576 - val_loss: 21.4126
Epoch 95/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 24.7782 - val_loss: 38.7811
Epoch 96/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 41.9597 - val_loss: 36.5568
Epoch 97/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 72.6424 - val_loss: 42.6189
Epoch 98/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 39.2912 - val_loss: 28.3757
Epoch 99/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 30.9735 - val_loss: 23.0405
Epoch 100/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 27.3346 - val_loss: 16.2927





<keras.src.callbacks.history.History at 0x24428576cd0>

验证集loss从294降到16.29

2.3.2 加入正则化惩罚项

model = tf.keras.Sequential()
model.add(layers.Dense(16, kernel_initializer='random_normal', kernel_regularizer=tf.keras.regularizers.l2(0.03)))
model.add(layers.Dense(32, kernel_initializer='random_normal', kernel_regularizer=tf.keras.regularizers.l2(0.03)))
model.add(layers.Dense(1, kernel_initializer='random_normal', kernel_regularizer=tf.keras.regularizers.l2(0.03)))
model.compile(optimizer=tf.keras.optimizers.SGD(0.001),
             loss='mean_squared_error')
model.fit(input_temperatures, labels, validation_split=0.25, epochs=100, batch_size=64)

Epoch 1/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 31ms/step - loss: 4329.9487 - val_loss: 2871.9626
Epoch 2/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 4307.8306 - val_loss: 2789.5034
Epoch 3/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 4179.3872 - val_loss: 2676.4529
Epoch 4/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 4030.3188 - val_loss: 2476.9949
Epoch 5/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 3222.8850 - val_loss: 2195.1760
Epoch 6/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 263.0312 - val_loss: 1695.6211
Epoch 7/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 68.0448 - val_loss: 1369.0785
Epoch 8/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 54.3513 - val_loss: 1032.3079
Epoch 9/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 39.0651 - val_loss: 831.0013
Epoch 10/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 37.9228 - val_loss: 672.9735
Epoch 11/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 35.9693 - val_loss: 450.8900
Epoch 12/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 54.7455 - val_loss: 481.6994
Epoch 13/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step - loss: 35.8592 - val_loss: 321.1757
Epoch 14/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 32.7608 - val_loss: 283.5308
Epoch 15/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 194.1068 - val_loss: 287.6764
Epoch 16/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 51.2676 - val_loss: 251.6622
Epoch 17/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 61.9756 - val_loss: 218.9387
Epoch 18/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 62.7216 - val_loss: 159.8604
Epoch 19/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 50.4749 - val_loss: 119.2429
Epoch 20/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 11ms/step - loss: 31.9662 - val_loss: 146.9869
Epoch 21/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 32.3689 - val_loss: 133.1835
Epoch 22/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 12ms/step - loss: 42.9962 - val_loss: 172.4340
Epoch 23/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 51.8156 - val_loss: 150.9075
Epoch 24/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 28.9330 - val_loss: 160.4308
Epoch 25/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 51.3750 - val_loss: 100.7717
Epoch 26/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 64.6685 - val_loss: 58.6360
Epoch 27/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step - loss: 28.2559 - val_loss: 77.3267
Epoch 28/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 30.2373 - val_loss: 74.5407
Epoch 29/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 39.5093 - val_loss: 44.6559
Epoch 30/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 44.0398 - val_loss: 64.8625
Epoch 31/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 59.9598 - val_loss: 52.8985
Epoch 32/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 37.3139 - val_loss: 44.8724
Epoch 33/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 32.5650 - val_loss: 56.2387
Epoch 34/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 11ms/step - loss: 32.8806 - val_loss: 54.2845
Epoch 35/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 29.2826 - val_loss: 71.4181
Epoch 36/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 43.0233 - val_loss: 52.3515
Epoch 37/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 53.5211 - val_loss: 63.1684
Epoch 38/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 35.4584 - val_loss: 46.4926
Epoch 39/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 40.3554 - val_loss: 41.1063
Epoch 40/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 24.9332 - val_loss: 57.6292
Epoch 41/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 33.9667 - val_loss: 33.8120
Epoch 42/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 26.6114 - val_loss: 43.4933
Epoch 43/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 30.9945 - val_loss: 39.8859
Epoch 44/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 31.4712 - val_loss: 44.3439
Epoch 45/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 30.3691 - val_loss: 28.9878
Epoch 46/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 31.7159 - val_loss: 29.2529
Epoch 47/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 29.2444 - val_loss: 52.1457
Epoch 48/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 45.8093 - val_loss: 29.0608
Epoch 49/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 37.3103 - val_loss: 25.5083
Epoch 50/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 24.3065 - val_loss: 22.4198
Epoch 51/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 26.5020 - val_loss: 75.6966
Epoch 52/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 108.3367 - val_loss: 26.7037
Epoch 53/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 24.1255 - val_loss: 26.3644
Epoch 54/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 26.4255 - val_loss: 21.8528
Epoch 55/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 30.5198 - val_loss: 30.5720
Epoch 56/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 75.6487 - val_loss: 22.3380
Epoch 57/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 27.7026 - val_loss: 47.1917
Epoch 58/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 91.9721 - val_loss: 20.5868
Epoch 59/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step - loss: 29.8157 - val_loss: 20.3234
Epoch 60/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step - loss: 25.5071 - val_loss: 21.0482
Epoch 61/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 30.9963 - val_loss: 26.6644
Epoch 62/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 36.8300 - val_loss: 53.5873
Epoch 63/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 121.0377 - val_loss: 23.9243
Epoch 64/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 25.9519 - val_loss: 38.1576
Epoch 65/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 58.6323 - val_loss: 25.9526
Epoch 66/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 33.2245 - val_loss: 28.1261
Epoch 67/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 32.8570 - val_loss: 26.6537
Epoch 68/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 31.3758 - val_loss: 24.2550


Epoch 69/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 33.5456 - val_loss: 26.4282
Epoch 70/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 24.6720 - val_loss: 25.8680
Epoch 71/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 39.6415 - val_loss: 30.0976
Epoch 72/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 3ms/step - loss: 31.5295 - val_loss: 22.4324
Epoch 73/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 31.7014 - val_loss: 28.8245
Epoch 74/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 28.5994 - val_loss: 19.5470
Epoch 75/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 25.5453 - val_loss: 20.3241
Epoch 76/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 25.2719 - val_loss: 30.4617
Epoch 77/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 37.2063 - val_loss: 28.1975
Epoch 78/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 27.9328 - val_loss: 25.9083
Epoch 79/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 26.2157 - val_loss: 21.1073
Epoch 80/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 26.5412 - val_loss: 26.4064
Epoch 81/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 32.7264 - val_loss: 22.4093
Epoch 82/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 39.1839 - val_loss: 28.2290
Epoch 83/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 25.7481 - val_loss: 23.3083
Epoch 84/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 30.1190 - val_loss: 29.3318
Epoch 85/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 29.4902 - val_loss: 21.6280
Epoch 86/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 34.3956 - val_loss: 18.1713
Epoch 87/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 29.2233 - val_loss: 19.9103
Epoch 88/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 38.5457 - val_loss: 22.7399
Epoch 89/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 27.8864 - val_loss: 26.5859
Epoch 90/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 33.6624 - val_loss: 24.8328
Epoch 91/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 38.2470 - val_loss: 17.1834
Epoch 92/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - loss: 30.2064 - val_loss: 25.3695
Epoch 93/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 32.6475 - val_loss: 46.5841
Epoch 94/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 9ms/step - loss: 58.2158 - val_loss: 23.7842
Epoch 95/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 22.8116 - val_loss: 37.8584
Epoch 96/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 89.6664 - val_loss: 27.8412
Epoch 97/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 33.9813 - val_loss: 29.5030
Epoch 98/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 9ms/step - loss: 27.7755 - val_loss: 41.9249
Epoch 99/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step - loss: 37.6500 - val_loss: 23.3591
Epoch 100/100
[1m5/5[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - loss: 30.2891 - val_loss: 20.3465





<keras.src.callbacks.history.History at 0x2442da3e610>

3.预测

3.1 预测结果

predict_result = model.predict(input_temperatures)
print("predict result shape ",predict_result.shape)
print(predict_result[0:5])

[1m11/11[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 2ms/step 
predict result shape  (348, 1)
[[49.465336]
 [51.12283 ]
 [47.591335]
 [48.294342]
 [45.045162]]

3.2 展示预测结果

#创建真实数据的展示
true_data = pd.DataFrame(data={'date': temp_days, 'actual': labels})

#解析日期，取出模型的预测值
predict_data = pd.DataFrame(data={'date': temp_days, 'prediction': predict_result.reshape(-1)})

plt.plot(true_data['date'], true_data['actual'], 'b-', label = 'actual')
plt.plot(predict_data['date'], predict_data['prediction'], 'ro', label = 'prediction')
plt.xticks(rotation = 60)
plt.legend()
plt.xlabel('Date')
plt.ylabel('Max Temperature')
plt.title('Actual And Predict Value')

Text(0.5, 1.0, 'Actual And Predict Value')

在这里插入图片描述

_feivirus_
关注
专栏目录
BP神经网络实现气温学习和预测代码及代码分析(python+tensorflow).zip
12-09
BP神经网络实现气温学习和预测代码及代码分析(python+tensflow)，用python+tensflow实现气温的预测，本代码包含数据集，不用重新下载，可以自行设置测试学习次数，误差范围，点开导包，运行即可使用，并且附上了因...
神经网络气温预测数据.rar
06-29
神经网络在气温预测中的应用是现代信息技术与气象科学深度融合的产物。神经网络作为一种模拟人脑神经元结构的计算模型，已经在许多领域展示了强大的预测能力，包括天气预报。本压缩包“神经网络气温预测数据.rar”...
基于Tensorflow的神经网络模型搭建——气温预测
struggle的博客
06-18 2608
基于Tensorflow的神经网络模型搭建
tensorflow对气温预测
qq_46026509的博客
04-01 687
数据表中数据维度: (348, 9)
搭建神经网络进行气温预测
不想随波逐流的大学生一枚，我们每个人都在努力奔跑
05-17 3185
TensorFlow搭建一个气温预测的小回归网络
唐宇迪TensorFlow2.0入门到实战：使用TensorFlow预测天气
x2yline的博客
07-02 728
数据表中year,month,day,week 分别表示具体的时间temp_2：前天的最高气温temp_1：昨天的最高气温average：在历史中，每年这一天的平均最高温度值actual：这就是我们的标签值了，当天的真实最高温度friend：这一列可能是凑热闹的，你的朋友猜测的可能值，咱们不管它就好。
基于TensorFlow的神经网络气温预测
sxmf566的博客
10-04 337
然后就可以用jupyter notebook导入数据做进一步分析。
【数值预测案例】(7) CNN-LSTM 混合神经网络气温预测，附TensorFlow完整代码
热门推荐
博观而约取，厚积而薄发
04-25 1万+
大家好，今天和各位分享一下如何使用 Tensorflow 构建 CNN卷积神经网络和 LSTM 循环神经网络相结合的混合神经网络模型，完成对多特征的时间序列预测。 本文预测模型的主要结构由 CNN 和 LSTM 神经网络构成。气温的特征数据具有空间依赖性。本文选择通过在模型前端使用CNN卷积神经网络提取特征之间的空间关系。同时，气温数据又具有明显的时间依赖性，因此在卷积神经网络后添加 LSTM 长短时记忆模型进行时序处理。 1. 获取数据集 数据集自取：https://download.csdn.ne
Tensorflow2.0 基于神经网络预测天气
dpoerpoiwetiop的博客
03-20 3539
标题tensorflow2.0的高级API使用起来非常方便 如果使用过程中需要查资料可以直接去官网查Api和使用教程 网址：https://www.tensorflow.org/api_docs/python/tf?hl=zh-cn 今天使用 全连接网络 就是数学中断 y=kx+b模式 常用参数 ： 1导入模块 import numpy as np import pandas as pd i...
使用BP神经网络实现时间序列预测
weixin_40963160的博客
07-15 100
在机器学习和人工智能领域，BP神经网络(Backpropagation Neural Network)作为一种经典的监督学习模型，广泛应用于时间序列预测任务中。本文将详细介绍如何使用Python语言和常见的机器学习库，构建和训练BP神经网络，以实现准确的时间序列预测。我们将深入探讨BP神经网络的工作原理、实现步骤以及如何优...
基于TensorFlow实现预测房价--回归问题
BoCong-Deng的博客
07-17 1547
分类问题的目标是预测输入数据点所对应的单一离散的标签。另一种常见的机器学习问题是回归问题，它预测一个连续值而不是离散的标签，例如，根据气象数据 预测明天的气温，或者根据软件说明书预测完成软件项目所需要的时间。
LSTM-morevalible.rar_LSTM_lstm 气温_lstm多预测_多变量预测_天气预测
07-13
LSTM是一种循环神经网络（RNN）的变体，旨在解决传统RNN在训练长期依赖性时遇到的梯度消失和爆炸问题。LSTM单元由三个门控机制组成：输入门、遗忘门和输出门。这些门控机制允许网络在处理序列时选择性地记住或忘记...
神经网络回归预测-气温数据集
06-05
在这个场景中，我们关注的是“神经网络回归预测-气温数据集”，这显然涉及到使用神经网络模型对气温进行时间序列预测。下面我们将深入探讨这个主题。 首先，神经网络是一种模仿人脑神经元结构的计算模型，它可以...
神经网络回归预测--气温数据集 - 副本.rar
06-21
在本案例中，我们关注的是如何运用神经网络对气温数据进行预测，这通常涉及时间序列分析，因为气温数据往往具有时间依赖性。`a.txt`可能是包含气温历史数据或模型训练结果的文本文件。 **神经网络基础** 神经网络是...
OpenSNN推文：神经网络(Neural Network)相关论文最新推荐（九月份）(二)
opensnn的博客
09-13 532
BP神经网络背景下应急通信网性能评价方法研究 论文链接：oalib 简介：  在列举和阐述应急通信网性能评价指标的基础上，本文首先利用熵权系数法确定了各指标权重，进而通过模糊综合评判法确定了评判权值，最后通过构造人工神经网络模型对应急通信网性能进行了评估。结果表明，利用人工神经网络进行评估的结果与目标值是一致的。 Keywords: BP神经网络，应急通信网，性能评价，熵权系数法 The Method of Feature Point Detection for Image Matching Bas
Python和R及MATLAB和C与Lua去相关生物医学图像处理和神经网络物理学及数学变换算法
最新发布
跨学科知识视角展现
09-16 823
1. 主成分分析降维2. 显微镜成像精度评估算法3. 脑电图磁共振成像降噪算法4. 图像颜色分离显现特征5. 球面转换:主成分分析和零相位分量分析6. 零相位分量分析和主成分分析平均互相关算法7. 图像白化8. 计算噪声协方差和绘制白化数据9. 高能物理分类器分离不同信号10. 白化变换优化批量归一化11. 去相关对比特征学习12. 生物图像去相关分析引擎
神经网络_使用tensorflow对mnist手写数字分类
研究源码的最底层,只持有正确的仓位
09-16 310
【代码】神经网络_使用tensorflow对mnist手写数字分类。
Transformer-Adaboost多输入单输出回归预测神经网络【MATLAB】
沅的博客
09-16 294
Transformer-Adaboost多输入单输出回归预测模型结合了Transformer架构和AdaBoost算法，用于处理时间序列数据或具有多个输入特征的回归任务。
卷积神经网络 tensorflow_神经网络原理演示动画合集
05-16
抱歉，我不太明白你的问题。你是想了解 TensorFlow 中的卷积神经网络以及神经网络原理的演示动画合集吗？如果是，我可以为您提供一些资源和链接。以下是一些相关的资源： 1. TensorFlow 官方文档：https://www.tensorflow.org/tutorials/ 2. TensorFlow 中文社区：https://www.tensorflow.org/community/translations/chinese 3. TensorFlow 中文教程：https://www.tensorflow.org/tutorials/?hl=zh-cn 4. TensorFlow 的 GitHub 页面：https://github.com/tensorflow/tensorflow 5. TensorFlow 的卷积神经网络教程：https://www.tensorflow.org/tutorials/images/cnn 6. 神经网络原理的演示动画合集：https://towardsdatascience.com/animated-intuition-for-cnns-and-rnns-8fbb9ccf6b17 希望这些资源能够帮助您更好地了解 TensorFlow 中的卷积神经网络和神经网络原理。
评论
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值