全连接神经网络实战

全连接神经网络实战

参考了书籍与博客，目标理解每句话的意思

import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import layers, losses

def load_dataset():
    dataset_path = keras.utils.get_file('auto-mpg.data','http://archive.ics.uci.edu/ml/machine-learning-databases/auto-mpg/auto-mpg.data')
    column_names = ['MPG','Cylinders','Displacement','Hoursepower','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()
    return dataset
dataset = load_dataset()
# 查看部分数据
dataset.head()

def preprocess_dataset(dataset):
    dataset = dataset.copy()
    #去除缺失值
    dataset = dataset.dropna()
    #将分类数字弹出
    origin = dataset.pop('Origin')
    #将分类数字化为虚拟变量
    dataset['USA'] = (origin==1)*1.0
    dataset['Europe'] = (origin==2)*1.0
    dataset['Japan'] = (origin == 3)*1.0
    #划分测试集与训练集
    train_dataset = dataset.sample(frac = 0.8,random_state = 0)
    test_dataset = dataset.drop(train_dataset.index)
    return train_dataset,test_dataset
train_dataset,test_dataset = preprocess_dataset(dataset)
train_dataset.head()
test_dataset.head()

#统计数据瞧瞧
sns_plot = sns.pairplot(train_dataset[['Cylinders','Displacement','Weight','MPG']],diag_kind ='kde')
#描述性统计分析
train_stats = train_dataset.describe()
train_stats.pop('MPG')
train_stats=train_stats.transpose()
train_stats

#数据格式化
def norm(x,train_stats):
    return (x-train_stats['mean'])/train_stats['std']

train_labels = train_dataset.pop('MPG')
test_labels = test_dataset.pop('MPG')
norm_train_dataset = norm(train_dataset,train_stats)
norm_test_dataset = norm(test_dataset,train_stats)
print(norm_train_dataset.shape,train_labels.shape)
print(norm_test_dataset.shape,test_labels.shape)

class Network(keras.Model):
    def __init__(self):
        super(Network,self).__init__()#调用父类（keras.Model的__init__()）
        #创建三层网络
        self.fc1 = layers.Dense(64,activation='relu')
        self.fc2 = layers.Dense(64,activation='relu')
        self.fc3 = layers.Dense(1)
    def call(self,inputs):
        x = self.fc1(inputs)
        x = self.fc2(x)
        x = self.fc3(x)
        return x

def build_model():
    model = Network()
    model.build(input_shape=(4,9)) #不要忘记人家继承了父类
    model.summary()
    return model

model = build_model()
optimizer = tf.keras.optimizers.RMSprop(0.001)
train_db = tf.data.Dataset.from_tensor_slices((norm_train_dataset.values,train_labels.values))#打标签
train_db = train_db.shuffle(100).batch(32)

def train(model,train_db,optimizer,norm_test_data,test_labels):
    train_mae_losses = []
    test_mae_losses = []
    for epoch in range(200):
        for step,(x,y) in enumerate(train_db):
            with tf.GradientTape() as Tape:
                out = model(x)
                loss = tf.reduce_mean(losses.MSE(y,out))
                mae_loss = tf.reduce_mean(losses.MAE(y,out))
            if step % 10 ==0:
                print(epoch,step,float(loss))
            grads =Tape.gradient(loss,model.trainable_variables)
            optimizer.apply_gradients(zip(grads,model.trainable_variables))
        train_mae_losses.append(float(mae_loss))
        out = model(tf.constant(norm_test_dataset.values))
        test_mae_losses.append(tf.reduce_mean(losses.MAE(test_labels,out)))
    return train_mae_losses, test_mae_losses


def plot(train_mae_losses, test_mae_losses):
    plt.figure()
    plt.xlabel('Epoch')
    plt.ylabel('MAE')
    plt.plot(train_mae_losses, label='Train')
    plt.plot(test_mae_losses, label='test')
    plt.legend()
    plt.legend()
    plt.show()

train_mae_losses,test_mae_losses = train(model,train_db,optimizer,norm_test_dataset,test_labels)
plot(train_mae_losses, test_mae_losses)