【机器学习】python借助pandas及scikit-learn使用三种方法分割训练集及测试集

分割方法可以分为分层抽样及随机抽样:
1.以下3种方法采取的是纯随机抽样的方法做划分。如果数据集足够庞大（特别是相较于属性的数量而言），这种方式通常不错。

首先加载原始数据：

import pandas as pd

def load_housing_data(housing_path=HOUSING_PATH):
    csv_path = os.path.join(housing_path, "housing.csv")
    return pd.read_csv(csv_path)
housing=load_housing_data()
1
2
3
4
5
6

1.使用纯概率法分割，缺点是新增训练数据后要重新分割：

import numpy as np

# For illustration only. Sklearn has train_test_split()
def split_train_test(data, test_ratio):
    #生成一个随机排列序列
    shuffled_indices = np.random.permutation(len(data))
    test_set_size = int(len(data) * test_ratio)
    test_indices = shuffled_indices[:test_set_size]
    train_indices = shuffled_indices[test_set_size:]
    #返回行列数据，这里只有行
    return data.iloc[train_indices], data.iloc[test_indices]
1
2
3
4
5
6
7
8
9
10
11

train_set, test_set = split_train_test(housing, 0.2)
len(train_set)

2.使用某一列哈希值做分割：

from zlib import crc32

def test_set_check(identifier, test_ratio):
    return crc32(np.int64(identifier)) & 0xffffffff < test_ratio * 2**32

def split_train_test_by_id(data, test_ratio, id_column):
    ids = data[id_column]
    in_test_set = ids.apply(lambda id_: test_set_check(id_, test_ratio))
    return data.loc[~in_test_set], data.loc[in_test_set]
1
2
3
4
5
6
7
8
9

#根据新的列值判断是否进入测试集：要求新增列具有唯一性，如没有可用索引列代替
housing_with_id = housing.reset_index()
train_set, test_set = split_train_test_by_id(housing_with_id, 0.2, “index”)

3.使用scikit-learn的自带分割函数：
#使用sklearn自带的方法

from sklearn.model_selection import train_test_split
train_set, test_set = train_test_split(housing, test_size=0.2, random_state=42)
1
2

下面以中等收入列进行分层抽样：
housing[“median_income”].hist()
housing[“median_income”].describe()

import numpy as np
#按照阈值箱转成分类号
housing["income_cat"] = pd.cut(housing["median_income"],
                               bins=[0., 1.5, 3.0, 4.5, 6., np.inf],
                               labels=[1, 2, 3, 4, 5])
#计算不同箱内所含的元素数目
housing["income_cat"].value_counts()
1
2
3
4
5
6
7

from sklearn.model_selection import StratifiedShuffleSplit
#不仅要进行随机打乱，最后选取的测试集和训练集的时候还要考虑样本类别的比例， 也就是我们说的分层抽样
split = StratifiedShuffleSplit(n_splits=1, test_size=0.2, random_state=42)
#指定需要分层的列,需要用for打开
for train_index, test_index in split.split(housing, housing["income_cat"]):
    strat_train_set = housing.loc[train_index]
    strat_test_set = housing.loc[test_index]
1
2
3
4
5
6
7

如此便完成了分层抽样。