mirror of
https://github.com/apachecn/ailearning.git
synced 2026-02-03 10:24:39 +08:00
t添加 实战项目_2_汽车燃油效率 文档和代码
This commit is contained in:
jiangzhonglian
903
docs/TensorFlow2.x/实战项目_2_汽车燃油效率.md
Normal file
903
docs/TensorFlow2.x/实战项目_2_汽车燃油效率.md
Normal file
@@ -0,0 +1,903 @@
|
||||
# 实战项目_2_汽车燃油效率
|
||||
|
||||
Note: 我们的 TensorFlow 社区翻译了这些文档。因为社区翻译是尽力而为， 所以无法保证它们是最准确的，并且反映了最新的
|
||||
[官方英文文档](https://www.tensorflow.org/?hl=en)。如果您有改进此翻译的建议， 请提交 pull request 到
|
||||
[tensorflow/docs](https://github.com/tensorflow/docs) GitHub 仓库。要志愿地撰写或者审核译文，请加入
|
||||
[docs-zh-cn@tensorflow.org Google Group](https://groups.google.com/a/tensorflow.org/forum/#!forum/docs-zh-cn)。
|
||||
|
||||
在 *回归 (regression)* 问题中，我们的目的是预测出如价格或概率这样连续值的输出。相对于*分类(classification)* 问题，*分类(classification)* 的目的是从一系列的分类出选择出一个分类 （如，给出一张包含苹果或橘子的图片，识别出图片中是哪种水果）。
|
||||
|
||||
本 notebook 使用经典的 [Auto MPG](https://archive.ics.uci.edu/ml/datasets/auto+mpg) 数据集，构建了一个用来预测70年代末到80年代初汽车燃油效率的模型。为了做到这一点，我们将为该模型提供许多那个时期的汽车描述。这个描述包含：气缸数，排量，马力以及重量。
|
||||
|
||||
本示例使用 `tf.keras` API，相关细节请参阅 [本指南](https://tensorflow.google.cn/guide/keras)。
|
||||
|
||||
|
||||
```python
|
||||
# 使用 seaborn 绘制矩阵图 (pairplot)
|
||||
!pip install seaborn
|
||||
```
|
||||
|
||||
Requirement already satisfied: seaborn in /usr/local/lib/python3.6/dist-packages (0.9.0)
|
||||
Requirement already satisfied: scipy>=0.14.0 in /usr/local/lib/python3.6/dist-packages (from seaborn) (1.3.1)
|
||||
Requirement already satisfied: numpy>=1.9.3 in /usr/local/lib/python3.6/dist-packages (from seaborn) (1.16.5)
|
||||
Requirement already satisfied: matplotlib>=1.4.3 in /usr/local/lib/python3.6/dist-packages (from seaborn) (3.0.3)
|
||||
Requirement already satisfied: pandas>=0.15.2 in /usr/local/lib/python3.6/dist-packages (from seaborn) (0.24.2)
|
||||
Requirement already satisfied: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=1.4.3->seaborn) (2.4.2)
|
||||
Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=1.4.3->seaborn) (1.1.0)
|
||||
Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=1.4.3->seaborn) (0.10.0)
|
||||
Requirement already satisfied: python-dateutil>=2.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib>=1.4.3->seaborn) (2.5.3)
|
||||
Requirement already satisfied: pytz>=2011k in /usr/local/lib/python3.6/dist-packages (from pandas>=0.15.2->seaborn) (2018.9)
|
||||
Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from kiwisolver>=1.0.1->matplotlib>=1.4.3->seaborn) (41.4.0)
|
||||
Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from cycler>=0.10->matplotlib>=1.4.3->seaborn) (1.12.0)
|
||||
|
||||
|
||||
|
||||
```python
|
||||
from __future__ import absolute_import, division, print_function, unicode_literals
|
||||
|
||||
import pathlib
|
||||
|
||||
import matplotlib.pyplot as plt
|
||||
import pandas as pd
|
||||
import seaborn as sns
|
||||
|
||||
try:
|
||||
# %tensorflow_version only exists in Colab.
|
||||
%tensorflow_version 2.x
|
||||
except Exception:
|
||||
pass
|
||||
import tensorflow as tf
|
||||
|
||||
from tensorflow import keras
|
||||
from tensorflow.keras import layers
|
||||
|
||||
print(tf.__version__)
|
||||
```
|
||||
|
||||
TensorFlow 2.x selected.
|
||||
2.0.0
|
||||
|
||||
|
||||
## Auto MPG 数据集
|
||||
|
||||
该数据集可以从 [UCI机器学习库](https://archive.ics.uci.edu/ml/) 中获取.
|
||||
|
||||
|
||||
|
||||
### 获取数据
|
||||
首先下载数据集。
|
||||
|
||||
|
||||
```python
|
||||
dataset_path = keras.utils.get_file("auto-mpg.data", "http://archive.ics.uci.edu/ml/machine-learning-databases/auto-mpg/auto-mpg.data")
|
||||
dataset_path
|
||||
```
|
||||
|
||||
Downloading data from http://archive.ics.uci.edu/ml/machine-learning-databases/auto-mpg/auto-mpg.data
|
||||
32768/30286 [================================] - 0s 4us/step
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
'/root/.keras/datasets/auto-mpg.data'
|
||||
|
||||
|
||||
|
||||
使用 pandas 导入数据集。
|
||||
|
||||
|
||||
```python
|
||||
column_names = ['MPG','Cylinders','Displacement','Horsepower','Weight',
|
||||
'Acceleration', 'Model Year', 'Origin']
|
||||
raw_dataset = pd.read_csv(dataset_path, names=column_names,
|
||||
na_values = "?", comment='\t',
|
||||
sep=" ", skipinitialspace=True)
|
||||
|
||||
dataset = raw_dataset.copy()
|
||||
dataset.tail()
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
||||
<div>
|
||||
<style scoped>
|
||||
.dataframe tbody tr th:only-of-type {
|
||||
vertical-align: middle;
|
||||
}
|
||||
|
||||
.dataframe tbody tr th {
|
||||
vertical-align: top;
|
||||
}
|
||||
|
||||
.dataframe thead th {
|
||||
text-align: right;
|
||||
}
|
||||
</style>
|
||||
<table border="1" class="dataframe">
|
||||
<thead>
|
||||
<tr style="text-align: right;">
|
||||
<th></th>
|
||||
<th>MPG</th>
|
||||
<th>Cylinders</th>
|
||||
<th>Displacement</th>
|
||||
<th>Horsepower</th>
|
||||
<th>Weight</th>
|
||||
<th>Acceleration</th>
|
||||
<th>Model Year</th>
|
||||
<th>Origin</th>
|
||||
</tr>
|
||||
</thead>
|
||||
<tbody>
|
||||
<tr>
|
||||
<th>393</th>
|
||||
<td>27.0</td>
|
||||
<td>4</td>
|
||||
<td>140.0</td>
|
||||
<td>86.0</td>
|
||||
<td>2790.0</td>
|
||||
<td>15.6</td>
|
||||
<td>82</td>
|
||||
<td>1</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>394</th>
|
||||
<td>44.0</td>
|
||||
<td>4</td>
|
||||
<td>97.0</td>
|
||||
<td>52.0</td>
|
||||
<td>2130.0</td>
|
||||
<td>24.6</td>
|
||||
<td>82</td>
|
||||
<td>2</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>395</th>
|
||||
<td>32.0</td>
|
||||
<td>4</td>
|
||||
<td>135.0</td>
|
||||
<td>84.0</td>
|
||||
<td>2295.0</td>
|
||||
<td>11.6</td>
|
||||
<td>82</td>
|
||||
<td>1</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>396</th>
|
||||
<td>28.0</td>
|
||||
<td>4</td>
|
||||
<td>120.0</td>
|
||||
<td>79.0</td>
|
||||
<td>2625.0</td>
|
||||
<td>18.6</td>
|
||||
<td>82</td>
|
||||
<td>1</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>397</th>
|
||||
<td>31.0</td>
|
||||
<td>4</td>
|
||||
<td>119.0</td>
|
||||
<td>82.0</td>
|
||||
<td>2720.0</td>
|
||||
<td>19.4</td>
|
||||
<td>82</td>
|
||||
<td>1</td>
|
||||
</tr>
|
||||
</tbody>
|
||||
</table>
|
||||
</div>
|
||||
|
||||
|
||||
|
||||
### 数据清洗
|
||||
|
||||
数据集中包括一些未知值。
|
||||
|
||||
|
||||
```python
|
||||
dataset.isna().sum()
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
||||
MPG 0
|
||||
Cylinders 0
|
||||
Displacement 0
|
||||
Horsepower 6
|
||||
Weight 0
|
||||
Acceleration 0
|
||||
Model Year 0
|
||||
Origin 0
|
||||
dtype: int64
|
||||
|
||||
|
||||
|
||||
为了保证这个初始示例的简单性，删除这些行。
|
||||
|
||||
|
||||
```python
|
||||
dataset = dataset.dropna()
|
||||
```
|
||||
|
||||
`"Origin"` 列实际上代表分类，而不仅仅是一个数字。所以把它转换为独热码 （one-hot）:
|
||||
|
||||
|
||||
```python
|
||||
origin = dataset.pop('Origin')
|
||||
```
|
||||
|
||||
|
||||
```python
|
||||
dataset['USA'] = (origin == 1)*1.0
|
||||
dataset['Europe'] = (origin == 2)*1.0
|
||||
dataset['Japan'] = (origin == 3)*1.0
|
||||
dataset.tail()
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
||||
<div>
|
||||
<style scoped>
|
||||
.dataframe tbody tr th:only-of-type {
|
||||
vertical-align: middle;
|
||||
}
|
||||
|
||||
.dataframe tbody tr th {
|
||||
vertical-align: top;
|
||||
}
|
||||
|
||||
.dataframe thead th {
|
||||
text-align: right;
|
||||
}
|
||||
</style>
|
||||
<table border="1" class="dataframe">
|
||||
<thead>
|
||||
<tr style="text-align: right;">
|
||||
<th></th>
|
||||
<th>MPG</th>
|
||||
<th>Cylinders</th>
|
||||
<th>Displacement</th>
|
||||
<th>Horsepower</th>
|
||||
<th>Weight</th>
|
||||
<th>Acceleration</th>
|
||||
<th>Model Year</th>
|
||||
<th>USA</th>
|
||||
<th>Europe</th>
|
||||
<th>Japan</th>
|
||||
</tr>
|
||||
</thead>
|
||||
<tbody>
|
||||
<tr>
|
||||
<th>393</th>
|
||||
<td>27.0</td>
|
||||
<td>4</td>
|
||||
<td>140.0</td>
|
||||
<td>86.0</td>
|
||||
<td>2790.0</td>
|
||||
<td>15.6</td>
|
||||
<td>82</td>
|
||||
<td>1.0</td>
|
||||
<td>0.0</td>
|
||||
<td>0.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>394</th>
|
||||
<td>44.0</td>
|
||||
<td>4</td>
|
||||
<td>97.0</td>
|
||||
<td>52.0</td>
|
||||
<td>2130.0</td>
|
||||
<td>24.6</td>
|
||||
<td>82</td>
|
||||
<td>0.0</td>
|
||||
<td>1.0</td>
|
||||
<td>0.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>395</th>
|
||||
<td>32.0</td>
|
||||
<td>4</td>
|
||||
<td>135.0</td>
|
||||
<td>84.0</td>
|
||||
<td>2295.0</td>
|
||||
<td>11.6</td>
|
||||
<td>82</td>
|
||||
<td>1.0</td>
|
||||
<td>0.0</td>
|
||||
<td>0.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>396</th>
|
||||
<td>28.0</td>
|
||||
<td>4</td>
|
||||
<td>120.0</td>
|
||||
<td>79.0</td>
|
||||
<td>2625.0</td>
|
||||
<td>18.6</td>
|
||||
<td>82</td>
|
||||
<td>1.0</td>
|
||||
<td>0.0</td>
|
||||
<td>0.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>397</th>
|
||||
<td>31.0</td>
|
||||
<td>4</td>
|
||||
<td>119.0</td>
|
||||
<td>82.0</td>
|
||||
<td>2720.0</td>
|
||||
<td>19.4</td>
|
||||
<td>82</td>
|
||||
<td>1.0</td>
|
||||
<td>0.0</td>
|
||||
<td>0.0</td>
|
||||
</tr>
|
||||
</tbody>
|
||||
</table>
|
||||
</div>
|
||||
|
||||
|
||||
|
||||
### 拆分训练数据集和测试数据集
|
||||
|
||||
现在需要将数据集拆分为一个训练数据集和一个测试数据集。
|
||||
|
||||
我们最后将使用测试数据集对模型进行评估。
|
||||
|
||||
|
||||
```python
|
||||
train_dataset = dataset.sample(frac=0.8,random_state=0)
|
||||
test_dataset = dataset.drop(train_dataset.index)
|
||||
```
|
||||
|
||||
### 数据检查
|
||||
|
||||
快速查看训练集中几对列的联合分布。
|
||||
|
||||
|
||||
```python
|
||||
sns.pairplot(train_dataset[["MPG", "Cylinders", "Displacement", "Weight"]], diag_kind="kde")
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
||||
<seaborn.axisgrid.PairGrid at 0x7f3d9243a6a0>
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
也可以查看总体的数据统计:
|
||||
|
||||
|
||||
```python
|
||||
train_stats = train_dataset.describe()
|
||||
train_stats.pop("MPG")
|
||||
train_stats = train_stats.transpose()
|
||||
train_stats
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
||||
<div>
|
||||
<style scoped>
|
||||
.dataframe tbody tr th:only-of-type {
|
||||
vertical-align: middle;
|
||||
}
|
||||
|
||||
.dataframe tbody tr th {
|
||||
vertical-align: top;
|
||||
}
|
||||
|
||||
.dataframe thead th {
|
||||
text-align: right;
|
||||
}
|
||||
</style>
|
||||
<table border="1" class="dataframe">
|
||||
<thead>
|
||||
<tr style="text-align: right;">
|
||||
<th></th>
|
||||
<th>count</th>
|
||||
<th>mean</th>
|
||||
<th>std</th>
|
||||
<th>min</th>
|
||||
<th>25%</th>
|
||||
<th>50%</th>
|
||||
<th>75%</th>
|
||||
<th>max</th>
|
||||
</tr>
|
||||
</thead>
|
||||
<tbody>
|
||||
<tr>
|
||||
<th>Cylinders</th>
|
||||
<td>314.0</td>
|
||||
<td>5.477707</td>
|
||||
<td>1.699788</td>
|
||||
<td>3.0</td>
|
||||
<td>4.00</td>
|
||||
<td>4.0</td>
|
||||
<td>8.00</td>
|
||||
<td>8.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>Displacement</th>
|
||||
<td>314.0</td>
|
||||
<td>195.318471</td>
|
||||
<td>104.331589</td>
|
||||
<td>68.0</td>
|
||||
<td>105.50</td>
|
||||
<td>151.0</td>
|
||||
<td>265.75</td>
|
||||
<td>455.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>Horsepower</th>
|
||||
<td>314.0</td>
|
||||
<td>104.869427</td>
|
||||
<td>38.096214</td>
|
||||
<td>46.0</td>
|
||||
<td>76.25</td>
|
||||
<td>94.5</td>
|
||||
<td>128.00</td>
|
||||
<td>225.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>Weight</th>
|
||||
<td>314.0</td>
|
||||
<td>2990.251592</td>
|
||||
<td>843.898596</td>
|
||||
<td>1649.0</td>
|
||||
<td>2256.50</td>
|
||||
<td>2822.5</td>
|
||||
<td>3608.00</td>
|
||||
<td>5140.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>Acceleration</th>
|
||||
<td>314.0</td>
|
||||
<td>15.559236</td>
|
||||
<td>2.789230</td>
|
||||
<td>8.0</td>
|
||||
<td>13.80</td>
|
||||
<td>15.5</td>
|
||||
<td>17.20</td>
|
||||
<td>24.8</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>Model Year</th>
|
||||
<td>314.0</td>
|
||||
<td>75.898089</td>
|
||||
<td>3.675642</td>
|
||||
<td>70.0</td>
|
||||
<td>73.00</td>
|
||||
<td>76.0</td>
|
||||
<td>79.00</td>
|
||||
<td>82.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>USA</th>
|
||||
<td>314.0</td>
|
||||
<td>0.624204</td>
|
||||
<td>0.485101</td>
|
||||
<td>0.0</td>
|
||||
<td>0.00</td>
|
||||
<td>1.0</td>
|
||||
<td>1.00</td>
|
||||
<td>1.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>Europe</th>
|
||||
<td>314.0</td>
|
||||
<td>0.178344</td>
|
||||
<td>0.383413</td>
|
||||
<td>0.0</td>
|
||||
<td>0.00</td>
|
||||
<td>0.0</td>
|
||||
<td>0.00</td>
|
||||
<td>1.0</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>Japan</th>
|
||||
<td>314.0</td>
|
||||
<td>0.197452</td>
|
||||
<td>0.398712</td>
|
||||
<td>0.0</td>
|
||||
<td>0.00</td>
|
||||
<td>0.0</td>
|
||||
<td>0.00</td>
|
||||
<td>1.0</td>
|
||||
</tr>
|
||||
</tbody>
|
||||
</table>
|
||||
</div>
|
||||
|
||||
|
||||
|
||||
### 从标签中分离特征
|
||||
|
||||
将特征值从目标值或者"标签"中分离。 这个标签是你使用训练模型进行预测的值。
|
||||
|
||||
|
||||
```python
|
||||
train_labels = train_dataset.pop('MPG')
|
||||
test_labels = test_dataset.pop('MPG')
|
||||
```
|
||||
|
||||
### 数据规范化
|
||||
|
||||
再次审视下上面的 `train_stats` 部分，并注意每个特征的范围有什么不同。
|
||||
|
||||
使用不同的尺度和范围对特征归一化是好的实践。尽管模型*可能* 在没有特征归一化的情况下收敛，它会使得模型训练更加复杂，并会造成生成的模型依赖输入所使用的单位选择。
|
||||
|
||||
注意：尽管我们仅仅从训练集中有意生成这些统计数据，但是这些统计信息也会用于归一化的测试数据集。我们需要这样做，将测试数据集放入到与已经训练过的模型相同的分布中。
|
||||
|
||||
|
||||
```python
|
||||
def norm(x):
|
||||
return (x - train_stats['mean']) / train_stats['std']
|
||||
normed_train_data = norm(train_dataset)
|
||||
normed_test_data = norm(test_dataset)
|
||||
```
|
||||
|
||||
我们将会使用这个已经归一化的数据来训练模型。
|
||||
|
||||
警告: 用于归一化输入的数据统计（均值和标准差）需要反馈给模型从而应用于任何其他数据，以及我们之前所获得独热码。这些数据包含测试数据集以及生产环境中所使用的实时数据。
|
||||
|
||||
## 模型
|
||||
|
||||
### 构建模型
|
||||
|
||||
让我们来构建我们自己的模型。这里，我们将会使用一个“顺序”模型，其中包含两个紧密相连的隐藏层，以及返回单个、连续值得输出层。模型的构建步骤包含于一个名叫 'build_model' 的函数中，稍后我们将会创建第二个模型。 两个密集连接的隐藏层。
|
||||
|
||||
|
||||
```python
|
||||
def build_model():
|
||||
model = keras.Sequential([
|
||||
layers.Dense(64, activation='relu', input_shape=[len(train_dataset.keys())]),
|
||||
layers.Dense(64, activation='relu'),
|
||||
layers.Dense(1)
|
||||
])
|
||||
|
||||
optimizer = tf.keras.optimizers.RMSprop(0.001)
|
||||
|
||||
model.compile(loss='mse',
|
||||
optimizer=optimizer,
|
||||
metrics=['mae', 'mse'])
|
||||
return model
|
||||
```
|
||||
|
||||
|
||||
```python
|
||||
model = build_model()
|
||||
```
|
||||
|
||||
### 检查模型
|
||||
|
||||
使用 `.summary` 方法来打印该模型的简单描述。
|
||||
|
||||
|
||||
```python
|
||||
model.summary()
|
||||
```
|
||||
|
||||
Model: "sequential"
|
||||
_________________________________________________________________
|
||||
Layer (type) Output Shape Param #
|
||||
=================================================================
|
||||
dense (Dense) (None, 64) 640
|
||||
_________________________________________________________________
|
||||
dense_1 (Dense) (None, 64) 4160
|
||||
_________________________________________________________________
|
||||
dense_2 (Dense) (None, 1) 65
|
||||
=================================================================
|
||||
Total params: 4,865
|
||||
Trainable params: 4,865
|
||||
Non-trainable params: 0
|
||||
_________________________________________________________________
|
||||
|
||||
|
||||
|
||||
现在试用下这个模型。从训练数据中批量获取‘10’条例子并对这些例子调用 `model.predict` 。
|
||||
|
||||
|
||||
|
||||
```python
|
||||
example_batch = normed_train_data[:10]
|
||||
example_result = model.predict(example_batch)
|
||||
example_result
|
||||
```
|
||||
|
||||
WARNING:tensorflow:Falling back from v2 loop because of error: Failed to find data adapter that can handle input: <class 'pandas.core.frame.DataFrame'>, <class 'NoneType'>
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
array([[ 0.37002987],
|
||||
[ 0.22292587],
|
||||
[ 0.7729857 ],
|
||||
[ 0.22504307],
|
||||
[-0.01411032],
|
||||
[ 0.25664118],
|
||||
[ 0.05221634],
|
||||
[-0.0256409 ],
|
||||
[ 0.23223272],
|
||||
[-0.00434934]], dtype=float32)
|
||||
|
||||
|
||||
|
||||
它似乎在工作，并产生了预期的形状和类型的结果
|
||||
|
||||
### 训练模型
|
||||
|
||||
对模型进行1000个周期的训练，并在 `history` 对象中记录训练和验证的准确性。
|
||||
|
||||
|
||||
```python
|
||||
# 通过为每个完成的时期打印一个点来显示训练进度
|
||||
class PrintDot(keras.callbacks.Callback):
|
||||
def on_epoch_end(self, epoch, logs):
|
||||
if epoch % 100 == 0: print('')
|
||||
print('.', end='')
|
||||
|
||||
EPOCHS = 1000
|
||||
|
||||
history = model.fit(
|
||||
normed_train_data, train_labels,
|
||||
epochs=EPOCHS, validation_split = 0.2, verbose=0,
|
||||
callbacks=[PrintDot()])
|
||||
```
|
||||
|
||||
WARNING:tensorflow:Falling back from v2 loop because of error: Failed to find data adapter that can handle input: <class 'pandas.core.frame.DataFrame'>, <class 'NoneType'>
|
||||
|
||||
....................................................................................................
|
||||
....................................................................................................
|
||||
....................................................................................................
|
||||
....................................................................................................
|
||||
....................................................................................................
|
||||
....................................................................................................
|
||||
....................................................................................................
|
||||
....................................................................................................
|
||||
....................................................................................................
|
||||
....................................................................................................
|
||||
|
||||
使用 `history` 对象中存储的统计信息可视化模型的训练进度。
|
||||
|
||||
|
||||
```python
|
||||
hist = pd.DataFrame(history.history)
|
||||
hist['epoch'] = history.epoch
|
||||
hist.tail()
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
||||
<div>
|
||||
<style scoped>
|
||||
.dataframe tbody tr th:only-of-type {
|
||||
vertical-align: middle;
|
||||
}
|
||||
|
||||
.dataframe tbody tr th {
|
||||
vertical-align: top;
|
||||
}
|
||||
|
||||
.dataframe thead th {
|
||||
text-align: right;
|
||||
}
|
||||
</style>
|
||||
<table border="1" class="dataframe">
|
||||
<thead>
|
||||
<tr style="text-align: right;">
|
||||
<th></th>
|
||||
<th>loss</th>
|
||||
<th>mae</th>
|
||||
<th>mse</th>
|
||||
<th>val_loss</th>
|
||||
<th>val_mae</th>
|
||||
<th>val_mse</th>
|
||||
<th>epoch</th>
|
||||
</tr>
|
||||
</thead>
|
||||
<tbody>
|
||||
<tr>
|
||||
<th>995</th>
|
||||
<td>2.278429</td>
|
||||
<td>0.968291</td>
|
||||
<td>2.278429</td>
|
||||
<td>8.645883</td>
|
||||
<td>2.228030</td>
|
||||
<td>8.645884</td>
|
||||
<td>995</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>996</th>
|
||||
<td>2.300897</td>
|
||||
<td>0.955693</td>
|
||||
<td>2.300897</td>
|
||||
<td>8.526561</td>
|
||||
<td>2.254299</td>
|
||||
<td>8.526561</td>
|
||||
<td>996</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>997</th>
|
||||
<td>2.302505</td>
|
||||
<td>0.937035</td>
|
||||
<td>2.302505</td>
|
||||
<td>8.662312</td>
|
||||
<td>2.204857</td>
|
||||
<td>8.662312</td>
|
||||
<td>997</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>998</th>
|
||||
<td>2.265367</td>
|
||||
<td>0.942647</td>
|
||||
<td>2.265367</td>
|
||||
<td>8.319109</td>
|
||||
<td>2.224138</td>
|
||||
<td>8.319109</td>
|
||||
<td>998</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<th>999</th>
|
||||
<td>2.224938</td>
|
||||
<td>0.985029</td>
|
||||
<td>2.224938</td>
|
||||
<td>8.404006</td>
|
||||
<td>2.241303</td>
|
||||
<td>8.404006</td>
|
||||
<td>999</td>
|
||||
</tr>
|
||||
</tbody>
|
||||
</table>
|
||||
</div>
|
||||
|
||||
|
||||
|
||||
|
||||
```python
|
||||
def plot_history(history):
|
||||
hist = pd.DataFrame(history.history)
|
||||
hist['epoch'] = history.epoch
|
||||
|
||||
plt.figure()
|
||||
plt.xlabel('Epoch')
|
||||
plt.ylabel('Mean Abs Error [MPG]')
|
||||
plt.plot(hist['epoch'], hist['mae'],
|
||||
label='Train Error')
|
||||
plt.plot(hist['epoch'], hist['val_mae'],
|
||||
label = 'Val Error')
|
||||
plt.ylim([0,5])
|
||||
plt.legend()
|
||||
|
||||
plt.figure()
|
||||
plt.xlabel('Epoch')
|
||||
plt.ylabel('Mean Square Error [$MPG^2$]')
|
||||
plt.plot(hist['epoch'], hist['mse'],
|
||||
label='Train Error')
|
||||
plt.plot(hist['epoch'], hist['val_mse'],
|
||||
label = 'Val Error')
|
||||
plt.ylim([0,20])
|
||||
plt.legend()
|
||||
plt.show()
|
||||
|
||||
|
||||
plot_history(history)
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
该图表显示在约100个 epochs 之后误差非但没有改进，反而出现恶化。 让我们更新 `model.fit` 调用，当验证值没有提高上是自动停止训练。
|
||||
我们将使用一个 *EarlyStopping callback* 来测试每个 epoch 的训练条件。如果经过一定数量的 epochs 后没有改进，则自动停止训练。
|
||||
|
||||
你可以从[这里](https://tensorflow.google.cn/versions/master/api_docs/python/tf/keras/callbacks/EarlyStopping)学习到更多的回调。
|
||||
|
||||
|
||||
```python
|
||||
model = build_model()
|
||||
|
||||
# patience 值用来检查改进 epochs 的数量
|
||||
early_stop = keras.callbacks.EarlyStopping(monitor='val_loss', patience=10)
|
||||
|
||||
history = model.fit(normed_train_data, train_labels, epochs=EPOCHS,
|
||||
validation_split = 0.2, verbose=0, callbacks=[early_stop, PrintDot()])
|
||||
|
||||
plot_history(history)
|
||||
```
|
||||
|
||||
WARNING:tensorflow:Falling back from v2 loop because of error: Failed to find data adapter that can handle input: <class 'pandas.core.frame.DataFrame'>, <class 'NoneType'>
|
||||
|
||||
........................................................
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
如图所示，验证集中的平均的误差通常在 +/- 2 MPG左右。 这个结果好么？ 我们将决定权留给你。
|
||||
|
||||
让我们看看通过使用 **测试集** 来泛化模型的效果如何，我们在训练模型时没有使用测试集。这告诉我们，当我们在现实世界中使用这个模型时，我们可以期望它预测得有多好。
|
||||
|
||||
|
||||
```python
|
||||
loss, mae, mse = model.evaluate(normed_test_data, test_labels, verbose=2)
|
||||
|
||||
print("Testing set Mean Abs Error: {:5.2f} MPG".format(mae))
|
||||
```
|
||||
|
||||
WARNING:tensorflow:Falling back from v2 loop because of error: Failed to find data adapter that can handle input: <class 'pandas.core.frame.DataFrame'>, <class 'NoneType'>
|
||||
78/78 - 0s - loss: 5.8737 - mae: 1.8844 - mse: 5.8737
|
||||
Testing set Mean Abs Error: 1.88 MPG
|
||||
|
||||
|
||||
### 做预测
|
||||
|
||||
最后，使用测试集中的数据预测 MPG 值:
|
||||
|
||||
|
||||
```python
|
||||
test_predictions = model.predict(normed_test_data).flatten()
|
||||
|
||||
plt.scatter(test_labels, test_predictions)
|
||||
plt.xlabel('True Values [MPG]')
|
||||
plt.ylabel('Predictions [MPG]')
|
||||
plt.axis('equal')
|
||||
plt.axis('square')
|
||||
plt.xlim([0,plt.xlim()[1]])
|
||||
plt.ylim([0,plt.ylim()[1]])
|
||||
_ = plt.plot([-100, 100], [-100, 100])
|
||||
|
||||
```
|
||||
|
||||
WARNING:tensorflow:Falling back from v2 loop because of error: Failed to find data adapter that can handle input: <class 'pandas.core.frame.DataFrame'>, <class 'NoneType'>
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
这看起来我们的模型预测得相当好。我们来看下误差分布。
|
||||
|
||||
|
||||
```python
|
||||
error = test_predictions - test_labels
|
||||
plt.hist(error, bins = 25)
|
||||
plt.xlabel("Prediction Error [MPG]")
|
||||
_ = plt.ylabel("Count")
|
||||
```
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
它不是完全的高斯分布，但我们可以推断出，这是因为样本的数量很小所导致的。
|
||||
|
||||
## 结论
|
||||
|
||||
本笔记本 (notebook) 介绍了一些处理回归问题的技术。
|
||||
|
||||
* 均方误差（MSE）是用于回归问题的常见损失函数（分类问题中使用不同的损失函数）。
|
||||
* 类似的，用于回归的评估指标与分类不同。 常见的回归指标是平均绝对误差（MAE）。
|
||||
* 当数字输入数据特征的值存在不同范围时，每个特征应独立缩放到相同范围。
|
||||
* 如果训练数据不多，一种方法是选择隐藏层较少的小网络，以避免过度拟合。
|
||||
* 早期停止是一种防止过度拟合的有效技术。
|
||||
281
src/py3.x/tensorflow2.x/text_regression.py
Normal file
281
src/py3.x/tensorflow2.x/text_regression.py
Normal file
@@ -0,0 +1,281 @@
|
||||
# -*- coding: utf-8 -*-
|
||||
"""text_regression.ipynb
|
||||
|
||||
Note: 我们的 TensorFlow 社区翻译了这些文档。因为社区翻译是尽力而为， 所以无法保证它们是最准确的，并且反映了最新的
|
||||
[官方英文文档](https://www.tensorflow.org/?hl=en)。如果您有改进此翻译的建议， 请提交 pull request 到
|
||||
[tensorflow/docs](https://github.com/tensorflow/docs) GitHub 仓库。要志愿地撰写或者审核译文，请加入
|
||||
[docs-zh-cn@tensorflow.org Google Group](https://groups.google.com/a/tensorflow.org/forum/#!forum/docs-zh-cn)。
|
||||
|
||||
在 *回归 (regression)* 问题中，我们的目的是预测出如价格或概率这样连续值的输出。相对于*分类(classification)* 问题，*分类(classification)* 的目的是从一系列的分类出选择出一个分类 （如，给出一张包含苹果或橘子的图片，识别出图片中是哪种水果）。
|
||||
|
||||
本 notebook 使用经典的 [Auto MPG](https://archive.ics.uci.edu/ml/datasets/auto+mpg) 数据集，构建了一个用来预测70年代末到80年代初汽车燃油效率的模型。为了做到这一点，我们将为该模型提供许多那个时期的汽车描述。这个描述包含：气缸数，排量，马力以及重量。
|
||||
|
||||
本示例使用 `tf.keras` API，相关细节请参阅 [本指南](https://tensorflow.google.cn/guide/keras)。
|
||||
"""
|
||||
|
||||
# 使用 seaborn 绘制矩阵图 (pairplot)
|
||||
!pip install seaborn
|
||||
|
||||
# Commented out IPython magic to ensure Python compatibility.
|
||||
from __future__ import absolute_import, division, print_function, unicode_literals
|
||||
|
||||
import pathlib
|
||||
|
||||
import matplotlib.pyplot as plt
|
||||
import pandas as pd
|
||||
import seaborn as sns
|
||||
|
||||
try:
|
||||
# %tensorflow_version only exists in Colab.
|
||||
# %tensorflow_version 2.x
|
||||
except Exception:
|
||||
pass
|
||||
import tensorflow as tf
|
||||
|
||||
from tensorflow import keras
|
||||
from tensorflow.keras import layers
|
||||
|
||||
print(tf.__version__)
|
||||
|
||||
"""## Auto MPG 数据集
|
||||
|
||||
该数据集可以从 [UCI机器学习库](https://archive.ics.uci.edu/ml/) 中获取.
|
||||
|
||||
### 获取数据
|
||||
首先下载数据集。
|
||||
"""
|
||||
|
||||
dataset_path = keras.utils.get_file("auto-mpg.data", "http://archive.ics.uci.edu/ml/machine-learning-databases/auto-mpg/auto-mpg.data")
|
||||
dataset_path
|
||||
|
||||
"""使用 pandas 导入数据集。"""
|
||||
|
||||
column_names = ['MPG','Cylinders','Displacement','Horsepower','Weight',
|
||||
'Acceleration', 'Model Year', 'Origin']
|
||||
raw_dataset = pd.read_csv(dataset_path, names=column_names,
|
||||
na_values = "?", comment='\t',
|
||||
sep=" ", skipinitialspace=True)
|
||||
|
||||
dataset = raw_dataset.copy()
|
||||
dataset.tail()
|
||||
|
||||
"""### 数据清洗
|
||||
|
||||
数据集中包括一些未知值。
|
||||
"""
|
||||
|
||||
dataset.isna().sum()
|
||||
|
||||
"""为了保证这个初始示例的简单性，删除这些行。"""
|
||||
|
||||
dataset = dataset.dropna()
|
||||
|
||||
"""`"Origin"` 列实际上代表分类，而不仅仅是一个数字。所以把它转换为独热码 （one-hot）:"""
|
||||
|
||||
origin = dataset.pop('Origin')
|
||||
|
||||
dataset['USA'] = (origin == 1)*1.0
|
||||
dataset['Europe'] = (origin == 2)*1.0
|
||||
dataset['Japan'] = (origin == 3)*1.0
|
||||
dataset.tail()
|
||||
|
||||
"""### 拆分训练数据集和测试数据集
|
||||
|
||||
现在需要将数据集拆分为一个训练数据集和一个测试数据集。
|
||||
|
||||
我们最后将使用测试数据集对模型进行评估。
|
||||
"""
|
||||
|
||||
train_dataset = dataset.sample(frac=0.8,random_state=0)
|
||||
test_dataset = dataset.drop(train_dataset.index)
|
||||
|
||||
"""### 数据检查
|
||||
|
||||
快速查看训练集中几对列的联合分布。
|
||||
"""
|
||||
|
||||
sns.pairplot(train_dataset[["MPG", "Cylinders", "Displacement", "Weight"]], diag_kind="kde")
|
||||
|
||||
"""也可以查看总体的数据统计:"""
|
||||
|
||||
train_stats = train_dataset.describe()
|
||||
train_stats.pop("MPG")
|
||||
train_stats = train_stats.transpose()
|
||||
train_stats
|
||||
|
||||
"""### 从标签中分离特征
|
||||
|
||||
将特征值从目标值或者"标签"中分离。 这个标签是你使用训练模型进行预测的值。
|
||||
"""
|
||||
|
||||
train_labels = train_dataset.pop('MPG')
|
||||
test_labels = test_dataset.pop('MPG')
|
||||
|
||||
"""### 数据规范化
|
||||
|
||||
再次审视下上面的 `train_stats` 部分，并注意每个特征的范围有什么不同。
|
||||
|
||||
使用不同的尺度和范围对特征归一化是好的实践。尽管模型*可能* 在没有特征归一化的情况下收敛，它会使得模型训练更加复杂，并会造成生成的模型依赖输入所使用的单位选择。
|
||||
|
||||
注意：尽管我们仅仅从训练集中有意生成这些统计数据，但是这些统计信息也会用于归一化的测试数据集。我们需要这样做，将测试数据集放入到与已经训练过的模型相同的分布中。
|
||||
"""
|
||||
|
||||
def norm(x):
|
||||
return (x - train_stats['mean']) / train_stats['std']
|
||||
normed_train_data = norm(train_dataset)
|
||||
normed_test_data = norm(test_dataset)
|
||||
|
||||
"""我们将会使用这个已经归一化的数据来训练模型。
|
||||
|
||||
警告: 用于归一化输入的数据统计（均值和标准差）需要反馈给模型从而应用于任何其他数据，以及我们之前所获得独热码。这些数据包含测试数据集以及生产环境中所使用的实时数据。
|
||||
|
||||
## 模型
|
||||
|
||||
### 构建模型
|
||||
|
||||
让我们来构建我们自己的模型。这里，我们将会使用一个“顺序”模型，其中包含两个紧密相连的隐藏层，以及返回单个、连续值得输出层。模型的构建步骤包含于一个名叫 'build_model' 的函数中，稍后我们将会创建第二个模型。 两个密集连接的隐藏层。
|
||||
"""
|
||||
|
||||
def build_model():
|
||||
model = keras.Sequential([
|
||||
layers.Dense(64, activation='relu', input_shape=[len(train_dataset.keys())]),
|
||||
layers.Dense(64, activation='relu'),
|
||||
layers.Dense(1)
|
||||
])
|
||||
|
||||
optimizer = tf.keras.optimizers.RMSprop(0.001)
|
||||
|
||||
model.compile(loss='mse',
|
||||
optimizer=optimizer,
|
||||
metrics=['mae', 'mse'])
|
||||
return model
|
||||
|
||||
model = build_model()
|
||||
|
||||
"""### 检查模型
|
||||
|
||||
使用 `.summary` 方法来打印该模型的简单描述。
|
||||
"""
|
||||
|
||||
model.summary()
|
||||
|
||||
"""现在试用下这个模型。从训练数据中批量获取‘10’条例子并对这些例子调用 `model.predict` 。"""
|
||||
|
||||
example_batch = normed_train_data[:10]
|
||||
example_result = model.predict(example_batch)
|
||||
example_result
|
||||
|
||||
"""它似乎在工作，并产生了预期的形状和类型的结果
|
||||
|
||||
### 训练模型
|
||||
|
||||
对模型进行1000个周期的训练，并在 `history` 对象中记录训练和验证的准确性。
|
||||
"""
|
||||
|
||||
# 通过为每个完成的时期打印一个点来显示训练进度
|
||||
class PrintDot(keras.callbacks.Callback):
|
||||
def on_epoch_end(self, epoch, logs):
|
||||
if epoch % 100 == 0: print('')
|
||||
print('.', end='')
|
||||
|
||||
EPOCHS = 1000
|
||||
|
||||
history = model.fit(
|
||||
normed_train_data, train_labels,
|
||||
epochs=EPOCHS, validation_split = 0.2, verbose=0,
|
||||
callbacks=[PrintDot()])
|
||||
|
||||
"""使用 `history` 对象中存储的统计信息可视化模型的训练进度。"""
|
||||
|
||||
hist = pd.DataFrame(history.history)
|
||||
hist['epoch'] = history.epoch
|
||||
hist.tail()
|
||||
|
||||
def plot_history(history):
|
||||
hist = pd.DataFrame(history.history)
|
||||
hist['epoch'] = history.epoch
|
||||
|
||||
plt.figure()
|
||||
plt.xlabel('Epoch')
|
||||
plt.ylabel('Mean Abs Error [MPG]')
|
||||
plt.plot(hist['epoch'], hist['mae'],
|
||||
label='Train Error')
|
||||
plt.plot(hist['epoch'], hist['val_mae'],
|
||||
label = 'Val Error')
|
||||
plt.ylim([0,5])
|
||||
plt.legend()
|
||||
|
||||
plt.figure()
|
||||
plt.xlabel('Epoch')
|
||||
plt.ylabel('Mean Square Error [$MPG^2$]')
|
||||
plt.plot(hist['epoch'], hist['mse'],
|
||||
label='Train Error')
|
||||
plt.plot(hist['epoch'], hist['val_mse'],
|
||||
label = 'Val Error')
|
||||
plt.ylim([0,20])
|
||||
plt.legend()
|
||||
plt.show()
|
||||
|
||||
|
||||
plot_history(history)
|
||||
|
||||
"""该图表显示在约100个 epochs 之后误差非但没有改进，反而出现恶化。 让我们更新 `model.fit` 调用，当验证值没有提高上是自动停止训练。
|
||||
我们将使用一个 *EarlyStopping callback* 来测试每个 epoch 的训练条件。如果经过一定数量的 epochs 后没有改进，则自动停止训练。
|
||||
|
||||
你可以从[这里](https://tensorflow.google.cn/versions/master/api_docs/python/tf/keras/callbacks/EarlyStopping)学习到更多的回调。
|
||||
"""
|
||||
|
||||
model = build_model()
|
||||
|
||||
# patience 值用来检查改进 epochs 的数量
|
||||
early_stop = keras.callbacks.EarlyStopping(monitor='val_loss', patience=10)
|
||||
|
||||
history = model.fit(normed_train_data, train_labels, epochs=EPOCHS,
|
||||
validation_split = 0.2, verbose=0, callbacks=[early_stop, PrintDot()])
|
||||
|
||||
plot_history(history)
|
||||
|
||||
"""如图所示，验证集中的平均的误差通常在 +/- 2 MPG左右。 这个结果好么？ 我们将决定权留给你。
|
||||
|
||||
让我们看看通过使用 **测试集** 来泛化模型的效果如何，我们在训练模型时没有使用测试集。这告诉我们，当我们在现实世界中使用这个模型时，我们可以期望它预测得有多好。
|
||||
"""
|
||||
|
||||
loss, mae, mse = model.evaluate(normed_test_data, test_labels, verbose=2)
|
||||
|
||||
print("Testing set Mean Abs Error: {:5.2f} MPG".format(mae))
|
||||
|
||||
"""### 做预测
|
||||
|
||||
最后，使用测试集中的数据预测 MPG 值:
|
||||
"""
|
||||
|
||||
test_predictions = model.predict(normed_test_data).flatten()
|
||||
|
||||
plt.scatter(test_labels, test_predictions)
|
||||
plt.xlabel('True Values [MPG]')
|
||||
plt.ylabel('Predictions [MPG]')
|
||||
plt.axis('equal')
|
||||
plt.axis('square')
|
||||
plt.xlim([0,plt.xlim()[1]])
|
||||
plt.ylim([0,plt.ylim()[1]])
|
||||
_ = plt.plot([-100, 100], [-100, 100])
|
||||
|
||||
"""这看起来我们的模型预测得相当好。我们来看下误差分布。"""
|
||||
|
||||
error = test_predictions - test_labels
|
||||
plt.hist(error, bins = 25)
|
||||
plt.xlabel("Prediction Error [MPG]")
|
||||
_ = plt.ylabel("Count")
|
||||
|
||||
"""它不是完全的高斯分布，但我们可以推断出，这是因为样本的数量很小所导致的。
|
||||
|
||||
## 结论
|
||||
|
||||
本笔记本 (notebook) 介绍了一些处理回归问题的技术。
|
||||
|
||||
* 均方误差（MSE）是用于回归问题的常见损失函数（分类问题中使用不同的损失函数）。
|
||||
* 类似的，用于回归的评估指标与分类不同。 常见的回归指标是平均绝对误差（MAE）。
|
||||
* 当数字输入数据特征的值存在不同范围时，每个特征应独立缩放到相同范围。
|
||||
* 如果训练数据不多，一种方法是选择隐藏层较少的小网络，以避免过度拟合。
|
||||
* 早期停止是一种防止过度拟合的有效技术。
|
||||
"""
|
||||
Reference in New Issue
Block a user