【入门教程】使用pytorch,sklearn,keras实现Boston房价预测

1.Pytorch对房价数据集的预测

1.1 加载原始数据以及所需要的包

–使用sklearn.datasets.load_boston对原始数据集进行加载

import torch.nn as nn
import numpy as np
import torch
from torch.utils.data import Dataset
from torch.utils.data import DataLoader
import torch.utils.data as Data
from sklearn.preprocessing import MinMaxScaler
#数据加载
from sklearn.datasets import load_boston
from sklearn.model_selection import train_test_split

对数据进行规范化，并划分训练集与测试集

data = load_boston()
x =data['data']
y = data['target']
y = y.reshape(-1,1)

# 数据规范化
mm_scale = MinMaxScaler()
x = mm_scale.fit_transform(x)
x_train,x_test,y_train,y_test = train_test_split(x,y,test_size=0.2)

1.2 搭建网络

构建最简单的网络：全连接，只有三层，13输入—>10个神经元的隐藏层—>10个神经元的隐藏层—>1输出。

#构造网络
class Model(torch.nn.Module):
    def __init__(self):
        super(Model, self).__init__()
        self.linear1 = torch.nn.Linear(13, 10)
        self.active2 = torch.nn.ReLU()
        self.linear3 = torch.nn.Linear(10,10)
        self.active4 = torch.nn.ReLU()
        self.linear5 = torch.nn.Linear(10,1)
    def forward(self, x):
        x = self.linear1(x)
        x = self.active2(x)
        x = self.linear3(x)
        x = self.active4(x)
        x = self.linear5(x)
        return x
         
model = Model()

1.3定义损失函数与优化器

nn.MSELoss()函数就是均方损失函数(y-y*)^2
优化器就是根据网络反向传播的梯度信息来更新网络的参数，以起到降低loss函数值的作用

criterion = nn.MSELoss()
optimizer = torch.optim.Adam(model.parameters(),lr=0.01)

1.4 模型训练

网络输入只能是张量，数据集是numpy.ndarray(可以使用train_set_input.type()查看数据类型)，需要使用torch.FloatTensor()进行类型转换。

x_train = torch.FloatTensor(x_train)
x_test = torch.FloatTensor(x_test)
y_train = torch.FloatTensor(y_train)
y_test = torch.FloatTensor(y_test)
#训练
max_epoch = 10000
for i in range(max_epoch):
    
    #前向传播
    y_pred = model(x_train)
    #计算loss
    loss = criterion(y_pred,y_train)
    #梯度清0
    optimizer.zero_grad()
    #反向传播
    loss.backward()
    #权重调整
    optimizer.step()

1.5 验证模型效果

output = model(x_test)
predict_list = output.detach().numpy()

print(np.hstack((y_test.numpy(),predict_list))[:10].T)
plt.plot(y_test.numpy(),color="red",linewidth=2)
plt.plot(predict_list,color="green",linewidth=2)
plt.show

预测结果展示：

--------------------------------------------------------------------------------------------------------------------------------

2.Sklearn对房价数据集的预测

2.1 加载原始数据以及所需要的包

–使用sklearn.datasets.load_boston对原始数据集进行加载

mport numpy as np
from sklearn.datasets import load_boston
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.preprocessing import MinMaxScaler
import matplotlib.pyplot as plt

data = load_boston()
X,y = data["data"],data["target"]
x_train,x_test,y_train,y_test = train_test_split(X,y,test_size=0.2,random_state=0)

2.2 数据预处理-归一化

MMS = MinMaxScaler()
x_train = MMS.fit_transform(x_train)
x_test = MMS.fit_transform(x_test)

2.3 模型的训练与预测

LR = LinearRegression()
LR.fit(x_train,y_train)

y_pred = LR.predict(x_test)
print(y_pred[:20],y_test[:20])
plt.plot(y_pred,color="red",linewidth=2)
plt.plot(y_test,color="green",linewidth=2)
plt.show

结果展示：

--------------------------------------------------------------------------------------------------------------------------------

3.Keras对房价数据集的预测

3.1 加载原始数据以及所需要的包

–使用sklearn.datasets.load_boston对原始数据集进行加载

import numpy as np
from sklearn.datasets import load_boston
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import MinMaxScaler
from keras.models import Sequential
from keras.layers import Dense
from keras import models
import matplotlib.pyplot as plt

data = load_boston()
X,y = data["data"],data["target"]
x_train,x_test,y_train,y_test = train_test_split(X,y,test_size=0.2,random_state=0)

3.2 数据预处理-归一化

MMS = MinMaxScaler()
x_train = MMS.fit_transform(x_train)
x_test = MMS.fit_transform(x_test)

3.3 搭建网络

构建最简单的网络：全连接，只有三层，13输入—>64个神经元的隐藏层—>64个神经元的隐藏层—>1输出。

model = models.Sequential()
model.add(Dense(64,activation='relu',input_dim=13))
model.add(Dense(64,activation='relu'))
model.add(Dense(1))

model.compile(loss='mse',optimizer='rmsprop',metrics=['mae'])
history = model.fit(x_train,y_train,batch_size=32,epochs=1000,verbose=0)

3.4 模型的预测

y_pred = model.predict(x_test)
print(np.hstack((y_pred,y_test.reshape(-1,1)))[:10])

plt.plot(y_pred,color="red",linewidth=2)
plt.plot(y_test,color="green",linewidth=2)
plt.show

预测结果展示：

以上，我们完成了使用pytorch，sklearn，keras分别对Boston房价数据集进行预测。