一起深度学习

预测房价：

import hashlib
import os
import tarfile
import zipfile
import requests

#@save
DATA_HUB = dict()
DATA_URL = 'http://d2l-data.s3-accelerate.amazonaws.com/'

#下载数据集
def download(name,cache_dir=os.path.join('..','data')):
    """下载一个DATA_HUB的文件，返回本地文件名"""
    assert name in DATA_HUB,f"{name} 不存在于 {DATA_HUB}"
    url, sha1_hash = DATA_HUB[name]
    #用于创建cahce_dir目录
    os.makedirs(cache_dir,exist_ok=True)
    fname = os.path.join(cache_dir,url.split('/')[-1])
    if os.path.exists(fname):
        sha1 = hashlib.sha1()
        with open(fname, 'rb') as f:
            while True:
                data = f.read(1048576)
                if not data:
                    break
                sha1.update(data)
        if sha1.hexdigest() == sha1_hash:
            return fname  # 命中缓存
    print(f'正在从{url}下载{fname}...')
    r = requests.get(url, stream=True, verify=True)
    with open(fname, 'wb') as f:
        f.write(r.content)
    return fname

def download_extract(name, folder=None):  #@save
    """下载并解压zip/tar文件"""
    fname = download(name)
    base_dir = os.path.dirname(fname)
    data_dir, ext = os.path.splitext(fname)
    if ext == '.zip':
        fp = zipfile.ZipFile(fname, 'r')
    elif ext in ('.tar', '.gz'):
        fp = tarfile.open(fname, 'r')
    else:
        assert False, '只有zip/tar文件可以被解压缩'
    fp.extractall(base_dir)
    return os.path.join(base_dir, folder) if folder else data_dir

def download_all():  #@save
    """下载DATA_HUB中的所有文件"""
    for name in DATA_HUB:
        download(name)


import numpy as np
import pandas as pd
import torch
from torch import nn
from d2l import torch as d2l
DATA_HUB['kaggle_house_train'] = (  #@save
    DATA_URL + 'kaggle_house_pred_train.csv',
    '585e9cc93e70b39160e7921475f9bcd7d31219ce')

DATA_HUB['kaggle_house_test'] = (  #@save
    DATA_URL + 'kaggle_house_pred_test.csv',
    'fa19780a7b011d9b009e8bff8e99922a8ee2eb90')
#加载数据
train_data = pd.read_csv(download('kaggle_house_train'))
test_data = pd.read_csv(download('kaggle_house_test'))

print(train_data.shape)
print(test_data.shape)

#选取前四行数据，以及第0，1，2，3列以及倒数1，2，3列
print(train_data.iloc[0:4,[0,1,2,3,-3,-2,-1]])

#将train_data 和test_data 整合在一起。
all_features = pd.concat((train_data.iloc[:, 1:-1], test_data.iloc[:, 1:]))
# print(all_features.shape)   #（2919，79）

#数据预处理
#为啥要进行数据预处理呢？
#标准化数据：1、方便优化  2、因为不知道有那些特征是相关的，不想让惩罚分配给一个特征的系数比分配给其他特征的系数更大，要雨露均沾。

#筛选出数据集中数据类型不是对象（非字符串）的特征列，并提取其索引，得到一个列表

numeric_features = all_features.dtypes[all_features.dtypes != 'object'].index
all_features[numeric_features] = all_features[numeric_features].apply(
    lambda x: (x - x.mean()) / (x.std())
)

#fillna(0)是将nan值替换为0
all_features[numeric_features] = all_features[numeric_features].fillna(0)
"""
| Feature_1 | Feature_2 | Feature_3 |
|-----------|-----------|-----------|
|    10     |    20     |   'A'     |
|    15     |    25     |   'B'     |
|    20     |    NaN    |   'A'     |
|    NaN    |    30     |   'B'     |
=>
| Feature_1 | Feature_2 | Feature_3 |
|-----------|-----------|-----------|
|   -1.0    |    -1.0   |   'A'     |
|    0.0    |     0.0   |   'B'     |
|    1.0    |    NaN    |   'A'     |
|    NaN    |     1.0   |   'B'     |
=>
| Feature_1 | Feature_2 | Feature_3 |
|-----------|-----------|-----------|
|   -1.0    |    -1.0   |   'A'     |
|    0.0    |     0.0   |   'B'     |
|    1.0    |     0.0   |   'A'     |
|    0.0    |     1.0   |   'B'     |

"""


all_features = pd.get_dummies(all_features,dummy_na=True)
all_features = all_features * 1
print(all_features.shape)

n_train = train_data.shape[0] #样本数  1460
train_features = torch.tensor(all_features[:n_train].values.astype(float), dtype=torch.float32)
test_features = torch.tensor(all_features[n_train:].values.astype(float), dtype=torch.float32)
train_labels = torch.tensor(
    train_data.SalePrice.values.astype(float).reshape(-1, 1), dtype=torch.float32)
# print(train_features.shape)

#训练
loss = nn.MSELoss()
in_features = train_features.shape[1] #取出输入特征值

def get_net():
    net = nn.Sequential(nn.Linear(in_features, 1))
    return net

def log_rmse(net, features, labels):
    # 为了在取对数时进一步稳定该值，将小于1的值设置为1
    clipped_preds = torch.clamp(net(features), 1, float('inf'))
    rmse = torch.sqrt(loss(torch.log(clipped_preds),
                           torch.log(labels)))
    return rmse.item()
def train(net, train_features, train_labels, test_features, test_labels,
          num_epochs, learning_rate, weight_decay, batch_size):
    train_ls, test_ls = [], []
    train_iter = d2l.load_array((train_features, train_labels), batch_size)
    # 这里使用的是Adam优化算法
    optimizer = torch.optim.Adam(net.parameters(),
                                 lr = learning_rate,
                                 weight_decay = weight_decay)
    for epoch in range(num_epochs):
        for X, y in train_iter:
            optimizer.zero_grad()
            l = loss(net(X), y)
            l.backward()
            optimizer.step()
        train_ls.append(log_rmse(net, train_features, train_labels))
        if test_labels is not None:
            test_ls.append(log_rmse(net, test_features, test_labels))
    return train_ls, test_ls
def get_k_fold_data(k, i, X, y):
    """
    实现k折交叉验证
    将x，y划分为训练集和验证集
    划分为 k 个子集，其中一个子集作为验证集，其余作为训练集。
    为啥要用 k折交叉验证呢？
    对于5次训练过程，每次所用到的训练集和验证集都不同，将训练得到的模型在验证集上进行评估，
    最后将这5次的验证结果进行平均，作为最终模型的性能评估指标。
    提高可靠性和稳定性。
    :param k:表示数据划分为几个子集，即k折交叉验证中的k值
    :param i:表示当前验证集的索引，取值范围为 0 到 k-1
    :param X:原始特征数据集
    :param y:原始标签数据集
    :return:
    """
    assert k > 1
    fold_size = X.shape[0] // k     #样本数量
    print("k:",k,"fold_size:",fold_size)
    X_train, y_train = None, None
    # k 有 5 个：0 ，1 ，2，3，4
    for j in range(k):
        idx = slice(j * fold_size, (j + 1) * fold_size)  #每次截取一部分
        X_part, y_part = X[idx, :], y[idx]
        if j == i:
            #验证集
            X_valid, y_valid = X_part, y_part
        elif X_train is None:
            #训练集为空，直接赋值
            X_train, y_train = X_part, y_part
        else:
            #训练集不为空，则进行拼接
            X_train = torch.cat([X_train, X_part], 0)
            y_train = torch.cat([y_train, y_part], 0)
    return X_train, y_train, X_valid, y_valid
def k_fold(k, X_train, y_train, num_epochs, learning_rate, weight_decay,
           batch_size):
    # print(X_train.shape) 1460
    train_l_sum, valid_l_sum = 0, 0

    for i in range(k):
        print('k_fold 的k：',k)
        # 5折，每次所用的训练集和验证集都会重新划分。
        data = get_k_fold_data(k, i, X_train, y_train)  #获取训练集和标签
        net = get_net() #获取神经网络
        train_ls, valid_ls = train(net, *data, num_epochs, learning_rate,
                                   weight_decay, batch_size)   #训练
        train_l_sum += train_ls[-1]
        valid_l_sum += valid_ls[-1]
        if i == 0:
            d2l.plot(list(range(1, num_epochs + 1)), [train_ls, valid_ls],
                     xlabel='epoch', ylabel='rmse', xlim=[1, num_epochs],
                     legend=['train', 'valid'], yscale='log')
        print(f'折{i + 1}，训练log rmse{float(train_ls[-1]):f}, '
              f'验证log rmse{float(valid_ls[-1]):f}')
    return train_l_sum / k, valid_l_sum / k

k, num_epochs, lr, batch_size = 5, 100, 30, 64
weight_decay = 0
train_l, valid_l = k_fold(k, train_features, train_labels, num_epochs, lr,
                          weight_decay, batch_size)
print(f'{k}-折验证: 平均训练log rmse: {float(train_l):f}, '
      f'平均验证log rmse: {float(valid_l):f}')