一个基于sklearn、pandas的小项目（个人兴趣）

机器学习项目

（关键词：sklearn、pandas、流水线pipeline、线性回归模型、支持向量机模型、随机森林模型）

数据集来自：https://www.kaggle.com/c/sberbank-russian-housing-market

数据集背景介绍：俄罗斯联邦储蓄银行正在挑战Kagglers开发使用广泛特征预测房地产价格的算法。竞争对手将依赖一个丰富的数据集，其中包括住房数据和宏观经济模式。准确的预测模型将使俄罗斯联邦储蓄银行能够在不确定的经济中为其客户提供更多的确定性。

import os
import pandas as pd
HOUSE_PATH = './datasets'
def load_housing_data(housing_path=HOUSE_PATH):
    csv_path = os.path.join(housing_path,'MoscowHouseTrain.csv')
    return pd.read_csv(csv_path)

housing = load_housing_data()
housing.head()

	id	timestamp	full_sq	life_sq	floor	max_floor	material	build_year	num_room	kitch_sq	...	cafe_count_5000_price_2500	cafe_count_5000_price_4000	cafe_count_5000_price_high	big_church_count_5000	church_count_5000	mosque_count_5000	leisure_count_5000	sport_count_5000	market_count_5000	price_doc
0	1	2011-08-20	43	27.0	4.0	NaN	NaN	NaN	NaN	NaN	...	9	4	0	13	22	1	0	52	4	5850000
1	2	2011-08-23	34	19.0	3.0	NaN	NaN	NaN	NaN	NaN	...	15	3	0	15	29	1	10	66	14	6000000
2	3	2011-08-27	43	29.0	2.0	NaN	NaN	NaN	NaN	NaN	...	10	3	0	11	27	0	4	67	10	5700000
3	4	2011-09-01	89	50.0	9.0	NaN	NaN	NaN	NaN	NaN	...	11	2	1	4	4	0	0	26	3	13100000
4	5	2011-09-05	77	77.0	4.0	NaN	NaN	NaN	NaN	NaN	...	319	108	17	135	236	2	91	195	14	16331452

5 rows × 292 columns

housingUse = housing[:10000]
housingUse.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 10000 entries, 0 to 9999
Columns: 292 entries, id to price_doc
dtypes: float64(119), int64(157), object(16)
memory usage: 22.3+ MB

housingUse.describe()

	id	full_sq	life_sq	floor	max_floor	material	build_year	num_room	kitch_sq	state	...	cafe_count_5000_price_2500	cafe_count_5000_price_4000	cafe_count_5000_price_high	big_church_count_5000	church_count_5000	mosque_count_5000	leisure_count_5000	sport_count_5000	market_count_5000	price_doc
count	10000.00000	10000.000000	7752.000000	9834.000000	494.000000	494.000000	437.000000	494.000000	494.000000	398.000000	...	10000.000000	10000.000000	10000.000000	10000.000000	10000.000000	10000.000000	10000.000000	10000.000000	10000.000000	1.000000e+04
mean	5002.24830	53.709900	34.181760	7.964308	11.921053	1.927126	1957.773455	1.935223	7.117409	2.419598	...	30.614500	10.164100	1.657500	14.497400	29.372800	0.488500	8.144400	52.529400	5.991500	6.400306e+06
std	2887.27684	58.306234	22.703808	5.383705	6.096573	1.472794	211.639811	0.774763	6.741713	0.752707	...	70.407453	27.145192	5.193601	27.874316	45.477441	0.606716	19.782004	44.945444	4.763559	4.376049e+06
min	1.00000	5.000000	0.000000	0.000000	0.000000	1.000000	0.000000	1.000000	0.000000	1.000000	...	0.000000	0.000000	0.000000	0.000000	1.000000	0.000000	0.000000	0.000000	0.000000	1.900000e+05
25%	2502.75000	38.000000	20.000000	4.000000	8.000000	1.000000	1965.000000	1.000000	5.000000	2.000000	...	2.000000	1.000000	0.000000	2.000000	8.000000	0.000000	0.000000	12.000000	1.000000	4.233045e+06
50%	5002.50000	47.000000	30.000000	7.000000	12.000000	1.000000	1976.000000	2.000000	7.000000	2.500000	...	8.000000	2.000000	0.000000	7.000000	16.000000	0.000000	2.000000	47.000000	5.000000	5.700000e+06
75%	7502.25000	61.250000	42.000000	11.000000	17.000000	2.000000	1997.000000	2.000000	9.000000	3.000000	...	20.000000	4.000000	1.000000	12.000000	28.000000	1.000000	6.000000	75.000000	10.000000	7.301552e+06
max	10002.00000	5326.000000	802.000000	44.000000	40.000000	6.000000	2015.000000	5.000000	123.000000	4.000000	...	377.000000	147.000000	30.000000	151.000000	250.000000	2.000000	105.000000	218.000000	20.000000	1.111111e+08

8 rows × 276 columns

%matplotlib inline
import matplotlib.pyplot as plt

fig, ax = plt.subplots(figsize=(60, 50))
housingUse.hist(ax=ax, bins=50, edgecolor="blue")

2、划分数据集

探索房屋目标销售价格作为分层抽样的可能性：

housingUse["price_doc"].hist(bins=70, edgecolor="blue")

housingUse["price_doc"].describe()

count    1.000000e+04
mean     6.400306e+06
std      4.376049e+06
min      1.900000e+05
25%      4.233045e+06
50%      5.700000e+06
75%      7.301552e+06
max      1.111111e+08
Name: price_doc, dtype: float64

import numpy as np

housingUse["income_cat"] = pd.cut(housingUse["price_doc"],
                               bins=[0., 4200000., 5700000., 7300000., 9000000., np.inf], 
                               labels=[1, 2, 3, 4, 5])
housingUse["income_cat"].hist(edgecolor="black")

# housingUse["income_cat"] = np.ceil(housingUse["price_doc"] / 100000000.)
# housingUse["income_cat"].where(housingUse["price_doc"] < 9, 9.0, inplace=True)
# housingUse["income_cat"].hist(edgecolor="black")

housingUse.head()

	id	timestamp	full_sq	life_sq	floor	max_floor	material	build_year	num_room	kitch_sq	...	cafe_count_5000_price_4000	cafe_count_5000_price_high	big_church_count_5000	church_count_5000	mosque_count_5000	leisure_count_5000	sport_count_5000	market_count_5000	price_doc	income_cat
0	1	2011-08-20	43	27.0	4.0	NaN	NaN	NaN	NaN	NaN	...	4	0	13	22	1	0	52	4	5850000	3
1	2	2011-08-23	34	19.0	3.0	NaN	NaN	NaN	NaN	NaN	...	3	0	15	29	1	10	66	14	6000000	3
2	3	2011-08-27	43	29.0	2.0	NaN	NaN	NaN	NaN	NaN	...	3	0	11	27	0	4	67	10	5700000	2
3	4	2011-09-01	89	50.0	9.0	NaN	NaN	NaN	NaN	NaN	...	2	1	4	4	0	0	26	3	13100000	5
4	5	2011-09-05	77	77.0	4.0	NaN	NaN	NaN	NaN	NaN	...	108	17	135	236	2	91	195	14	16331452	5

5 rows × 293 columns

housingUse["income_cat"].value_counts()/len(housing)

2    0.085754
1    0.080995
3    0.079387
5    0.048013
4    0.034032
Name: income_cat, dtype: float64

# 基于目标售价的分层抽样
from sklearn.model_selection import StratifiedShuffleSplit

split = StratifiedShuffleSplit(n_splits=1, test_size=0.2, random_state=42)
for train_index, test_index in split.split(housingUse, housingUse["income_cat"]):
    strat_train_set = housingUse.loc[train_index]
    strat_test_set = housingUse.loc[test_index]

# 完事以后，删除用于抽样而增加的那一列的信息
strat_train_set.drop('income_cat',axis=1,inplace=True)  
strat_test_set.drop('income_cat',axis=1,inplace=True)

print("训练集大小：" + str(strat_train_set.shape) + "\n测试集大小：" + str(strat_test_set.shape))

训练集大小：(8000, 292)
测试集大小：(2000, 292)

3、探索数据集相关性

housing = strat_train_set.copy()

corr_matrix = housing.corr()

corr_matrix

	id	full_sq	life_sq	floor	max_floor	material	build_year	num_room	kitch_sq	state	...	cafe_count_5000_price_2500	cafe_count_5000_price_4000	cafe_count_5000_price_high	big_church_count_5000	church_count_5000	mosque_count_5000	leisure_count_5000	sport_count_5000	market_count_5000	price_doc
id	1.000000	0.012074	0.121534	0.087292	-0.013266	-0.003899	-0.060945	-0.043295	-0.056324	-0.016977	...	-0.042916	-0.033283	-0.023342	-0.059320	-0.061332	-0.002739	-0.047711	-0.163295	-0.185277	0.012557
full_sq	0.012074	1.000000	0.208281	0.076964	0.148558	-0.085975	-0.157111	0.700592	0.232639	-0.067896	...	0.010269	0.009751	0.012328	0.002156	0.004670	0.004773	0.004127	-0.011730	-0.031863	0.198563
life_sq	0.121534	0.208281	1.000000	0.132810	0.047904	-0.051019	-0.042960	0.395729	0.143683	-0.020084	...	0.036823	0.041765	0.048429	0.019060	0.017277	-0.004664	0.022581	-0.069183	-0.138471	0.360761
floor	0.087292	0.076964	0.132810	1.000000	0.464896	-0.077248	-0.146436	-0.033062	0.040441	-0.012902	...	-0.048963	-0.047175	-0.038238	-0.058347	-0.059083	-0.003731	-0.059940	-0.129194	-0.147641	0.138167
max_floor	-0.013266	0.148558	0.047904	0.464896	1.000000	0.033218	0.006264	-0.043865	0.124776	-0.049004	...	-0.226110	-0.223682	-0.224157	-0.194559	-0.214707	-0.194212	-0.225493	-0.241819	-0.207358	0.112951
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
mosque_count_5000	-0.002739	0.004773	-0.004664	-0.003731	-0.194212	-0.009277	0.065323	0.128494	-0.064821	0.177200	...	0.485745	0.453952	0.444877	0.448511	0.505794	1.000000	0.456811	0.439594	0.170313	0.129135
leisure_count_5000	-0.047711	0.004127	0.022581	-0.059940	-0.225493	0.068069	0.023510	0.151868	-0.007931	0.118095	...	0.985805	0.974862	0.945952	0.965860	0.973137	0.456811	1.000000	0.806361	0.460463	0.179983
sport_count_5000	-0.163295	-0.011730	-0.069183	-0.129194	-0.241819	0.122477	0.077763	0.155859	-0.021454	0.223212	...	0.820315	0.771293	0.743241	0.822253	0.845495	0.439594	0.806361	1.000000	0.729183	0.263404
market_count_5000	-0.185277	-0.031863	-0.138471	-0.147641	-0.207358	0.053439	0.098520	0.062009	-0.014324	0.296082	...	0.429567	0.375837	0.345498	0.479993	0.512822	0.170313	0.460463	0.729183	1.000000	0.172720
price_doc	0.012557	0.198563	0.360761	0.138167	0.112951	-0.049962	-0.020732	0.495512	0.125347	0.170591	...	0.203641	0.184150	0.188428	0.182466	0.193230	0.129135	0.179983	0.263404	0.172720	1.000000

276 rows × 276 columns

# 跟售价最相关的特征
related_features = corr_matrix["price_doc"].sort_values(ascending=False)

print(related_features[0:11])

price_doc              1.000000
num_room               0.495512
life_sq                0.360761
sport_count_3000       0.289984
sport_count_2000       0.281693
trc_count_5000         0.267114
sport_count_5000       0.263404
sport_count_1500       0.254639
trc_count_3000         0.254040
sport_objects_raion    0.238254
trc_sqm_5000           0.233633
Name: price_doc, dtype: float64

top_10_features = list(related_features.index[0:11]) # 去掉特征A自身
print(top_10_features)


from pandas.plotting import scatter_matrix

scatter_matrix(housing[top_10_features], figsize=(24, 16))
# plt.scatter(housing[top_10_features.index[0]], housing['price_doc'], s=5)
# for f in top_10_features.index[1:]:
#     plt.scatter(housing[f], housing['price_doc'], s=5)

# plt.xlabel("Top 10 Features")
# plt.ylabel("price_doc")

# 房屋数似乎是跟售价最相关的特征
housing.plot(kind="scatter", x="num_room", y="price_doc", alpha=0.1)

<AxesSubplot:xlabel='num_room', ylabel='price_doc'>

分离训练集和标签

# 分离训练集和标签
housing_train = strat_train_set.drop("price_doc", axis=1)  #拷贝了一份，而不会改变原始数据。如果inplace=true那就会改变
housing_train_labels = strat_train_set["price_doc"].copy()

print("机器学习训练集：" + str(housing_train.shape) + "\n标签：" + str(housing_train_labels.shape))

机器学习训练集：(8000, 291)
标签：(8000,)

pd.isna(housing_train["max_floor"]).sum()
housing_train.dtypes

id                      int64
timestamp              object
full_sq                 int64
life_sq               float64
floor                 float64
                       ...   
church_count_5000       int64
mosque_count_5000       int64
leisure_count_5000      int64
sport_count_5000        int64
market_count_5000       int64
Length: 291, dtype: object

# # 获取所有文本类型特征
# text_features = housing_train.select_dtypes(include=[np.object])

# # 获取所有数值类型特征
# numeric_features = housing_train.select_dtypes(include=[np.number])

# # 输出结果
# print("文本类型特征：\n", text_features)
# print("数值类型特征：\n", numeric_features)


#获取所有文本类型特征的名字列表
text_feature_names = list(housing_train.select_dtypes(include=[np.object]).columns)

# 获取所有数值类型特征的名字列表
numeric_feature_names = list(housing_train.select_dtypes(include=[np.number]).columns)

# 输出结果
print("文本类型特征的名字列表：", text_feature_names)
print("数值类型特征的名字列表：", numeric_feature_names)

文本类型特征的名字列表： ['timestamp', 'product_type', 'sub_area', 'culture_objects_top_25', 'thermal_power_plant_raion', 'incineration_raion', 'oil_chemistry_raion', 'radiation_raion', 'railroad_terminal_raion', 'big_market_raion', 'nuclear_reactor_raion', 'detention_facility_raion', 'water_1line', 'big_road1_1line', 'railroad_1line', 'ecology']
数值类型特征的名字列表： ['id', 'full_sq', 'life_sq', 'floor', 'max_floor', 'material', 'build_year', 'num_room', 'kitch_sq', 'state', 'area_m', 'raion_popul', 'green_zone_part', 'indust_part', 'children_preschool', 'preschool_quota', 'preschool_education_centers_raion', 'children_school', 'school_quota', 'school_education_centers_raion', 'school_education_centers_top_20_raion', 'hospital_beds_raion', 'healthcare_centers_raion', 'university_top_20_raion', 'sport_objects_raion', 'additional_education_raion', 'culture_objects_top_25_raion', 'shopping_centers_raion', 'office_raion', 'full_all', 'male_f', 'female_f', 'young_all', 'young_male', 'young_female', 'work_all', 'work_male', 'work_female', 'ekder_all', 'ekder_male', 'ekder_female', '0_6_all', '0_6_male', '0_6_female', '7_14_all', '7_14_male', '7_14_female', '0_17_all', '0_17_male', '0_17_female', '16_29_all', '16_29_male', '16_29_female', '0_13_all', '0_13_male', '0_13_female', 'raion_build_count_with_material_info', 'build_count_block', 'build_count_wood', 'build_count_frame', 'build_count_brick', 'build_count_monolith', 'build_count_panel', 'build_count_foam', 'build_count_slag', 'build_count_mix', 'raion_build_count_with_builddate_info', 'build_count_before_1920', 'build_count_1921-1945', 'build_count_1946-1970', 'build_count_1971-1995', 'build_count_after_1995', 'ID_metro', 'metro_min_avto', 'metro_km_avto', 'metro_min_walk', 'metro_km_walk', 'kindergarten_km', 'school_km', 'park_km', 'green_zone_km', 'industrial_km', 'water_treatment_km', 'cemetery_km', 'incineration_km', 'railroad_station_walk_km', 'railroad_station_walk_min', 'ID_railroad_station_walk', 'railroad_station_avto_km', 'railroad_station_avto_min', 'ID_railroad_station_avto', 'public_transport_station_km', 'public_transport_station_min_walk', 'water_km', 'mkad_km', 'ttk_km', 'sadovoe_km', 'bulvar_ring_km', 'kremlin_km', 'big_road1_km', 'ID_big_road1', 'big_road2_km', 'ID_big_road2', 'railroad_km', 'zd_vokzaly_avto_km', 'ID_railroad_terminal', 'bus_terminal_avto_km', 'ID_bus_terminal', 'oil_chemistry_km', 'nuclear_reactor_km', 'radiation_km', 'power_transmission_line_km', 'thermal_power_plant_km', 'ts_km', 'big_market_km', 'market_shop_km', 'fitness_km', 'swim_pool_km', 'ice_rink_km', 'stadium_km', 'basketball_km', 'hospice_morgue_km', 'detention_facility_km', 'public_healthcare_km', 'university_km', 'workplaces_km', 'shopping_centers_km', 'office_km', 'additional_education_km', 'preschool_km', 'big_church_km', 'church_synagogue_km', 'mosque_km', 'theater_km', 'museum_km', 'exhibition_km', 'catering_km', 'green_part_500', 'prom_part_500', 'office_count_500', 'office_sqm_500', 'trc_count_500', 'trc_sqm_500', 'cafe_count_500', 'cafe_sum_500_min_price_avg', 'cafe_sum_500_max_price_avg', 'cafe_avg_price_500', 'cafe_count_500_na_price', 'cafe_count_500_price_500', 'cafe_count_500_price_1000', 'cafe_count_500_price_1500', 'cafe_count_500_price_2500', 'cafe_count_500_price_4000', 'cafe_count_500_price_high', 'big_church_count_500', 'church_count_500', 'mosque_count_500', 'leisure_count_500', 'sport_count_500', 'market_count_500', 'green_part_1000', 'prom_part_1000', 'office_count_1000', 'office_sqm_1000', 'trc_count_1000', 'trc_sqm_1000', 'cafe_count_1000', 'cafe_sum_1000_min_price_avg', 'cafe_sum_1000_max_price_avg', 'cafe_avg_price_1000', 'cafe_count_1000_na_price', 'cafe_count_1000_price_500', 'cafe_count_1000_price_1000', 'cafe_count_1000_price_1500', 'cafe_count_1000_price_2500', 'cafe_count_1000_price_4000', 'cafe_count_1000_price_high', 'big_church_count_1000', 'church_count_1000', 'mosque_count_1000', 'leisure_count_1000', 'sport_count_1000', 'market_count_1000', 'green_part_1500', 'prom_part_1500', 'office_count_1500', 'office_sqm_1500', 'trc_count_1500', 'trc_sqm_1500', 'cafe_count_1500', 'cafe_sum_1500_min_price_avg', 'cafe_sum_1500_max_price_avg', 'cafe_avg_price_1500', 'cafe_count_1500_na_price', 'cafe_count_1500_price_500', 'cafe_count_1500_price_1000', 'cafe_count_1500_price_1500', 'cafe_count_1500_price_2500', 'cafe_count_1500_price_4000', 'cafe_count_1500_price_high', 'big_church_count_1500', 'church_count_1500', 'mosque_count_1500', 'leisure_count_1500', 'sport_count_1500', 'market_count_1500', 'green_part_2000', 'prom_part_2000', 'office_count_2000', 'office_sqm_2000', 'trc_count_2000', 'trc_sqm_2000', 'cafe_count_2000', 'cafe_sum_2000_min_price_avg', 'cafe_sum_2000_max_price_avg', 'cafe_avg_price_2000', 'cafe_count_2000_na_price', 'cafe_count_2000_price_500', 'cafe_count_2000_price_1000', 'cafe_count_2000_price_1500', 'cafe_count_2000_price_2500', 'cafe_count_2000_price_4000', 'cafe_count_2000_price_high', 'big_church_count_2000', 'church_count_2000', 'mosque_count_2000', 'leisure_count_2000', 'sport_count_2000', 'market_count_2000', 'green_part_3000', 'prom_part_3000', 'office_count_3000', 'office_sqm_3000', 'trc_count_3000', 'trc_sqm_3000', 'cafe_count_3000', 'cafe_sum_3000_min_price_avg', 'cafe_sum_3000_max_price_avg', 'cafe_avg_price_3000', 'cafe_count_3000_na_price', 'cafe_count_3000_price_500', 'cafe_count_3000_price_1000', 'cafe_count_3000_price_1500', 'cafe_count_3000_price_2500', 'cafe_count_3000_price_4000', 'cafe_count_3000_price_high', 'big_church_count_3000', 'church_count_3000', 'mosque_count_3000', 'leisure_count_3000', 'sport_count_3000', 'market_count_3000', 'green_part_5000', 'prom_part_5000', 'office_count_5000', 'office_sqm_5000', 'trc_count_5000', 'trc_sqm_5000', 'cafe_count_5000', 'cafe_sum_5000_min_price_avg', 'cafe_sum_5000_max_price_avg', 'cafe_avg_price_5000', 'cafe_count_5000_na_price', 'cafe_count_5000_price_500', 'cafe_count_5000_price_1000', 'cafe_count_5000_price_1500', 'cafe_count_5000_price_2500', 'cafe_count_5000_price_4000', 'cafe_count_5000_price_high', 'big_church_count_5000', 'church_count_5000', 'mosque_count_5000', 'leisure_count_5000', 'sport_count_5000', 'market_count_5000']

4、开始处理数据

# 数据处理前的准备
from sklearn.base import BaseEstimator, TransformerMixin

# scikit-learn暂时不支持DataFrame，因此我们需要分别选择数值型和枚举型
class DataFrameSelector(BaseEstimator, TransformerMixin):
    def __init__(self, attribute_names):
        self.attribute_names = attribute_names
    def fit(self, X, y=None):
        return self
    def transform(self, X):
        return X[self.attribute_names].values
    
num_pipeline = Pipeline([
    ('selector', DataFrameSelector(numeric_feature_names)),
    ('imputer', SimpleImputer(strategy="median")),
    ('std_scaler', StandardScaler())
])

num_pipeline = Pipeline([
    ('selector', DataFrameSelector(numeric_feature_names)),
    ('imputer', SimpleImputer(strategy="median")),
    ('std_scaler', StandardScaler())
])

housing_train_num = housing_train.drop(housing_train[text_feature_names], axis=1)
housing_train_num.describe()

	id	full_sq	life_sq	floor	max_floor	material	build_year	num_room	kitch_sq	state	...	cafe_count_5000_price_1500	cafe_count_5000_price_2500	cafe_count_5000_price_4000	cafe_count_5000_price_high	big_church_count_5000	church_count_5000	mosque_count_5000	leisure_count_5000	sport_count_5000	market_count_5000
count	8000.000000	8000.000000	6220.000000	7873.000000	390.000000	390.000000	348.000000	390.000000	390.000000	312.000000	...	8000.000000	8000.000000	8000.000000	8000.000000	8000.000000	8000.000000	8000.000000	8000.000000	8000.000000	8000.000000
mean	4981.756625	53.796250	34.161415	8.002794	12.061538	1.892308	1956.942529	1.933333	7.156410	2.448718	...	61.526625	30.692375	10.188875	1.665625	14.550750	29.439000	0.487750	8.170375	52.779125	6.010375
std	2899.039994	64.166145	23.385401	5.412505	6.069970	1.435393	212.097490	0.785908	7.406913	0.733181	...	119.186798	70.607171	27.197218	5.218378	27.970448	45.681075	0.608602	19.862469	44.979570	4.771961
min	1.000000	5.000000	0.000000	0.000000	0.000000	1.000000	0.000000	1.000000	0.000000	1.000000	...	0.000000	0.000000	0.000000	0.000000	0.000000	2.000000	0.000000	0.000000	0.000000	0.000000
25%	2470.750000	38.000000	20.000000	4.000000	9.000000	1.000000	1965.000000	1.000000	5.000000	2.000000	...	8.000000	2.000000	1.000000	0.000000	2.000000	8.000000	0.000000	0.000000	12.000000	1.000000
50%	4982.500000	46.000000	30.000000	7.000000	12.000000	1.000000	1975.000000	2.000000	7.000000	3.000000	...	24.000000	9.000000	2.000000	0.000000	7.000000	16.000000	0.000000	2.000000	47.000000	5.000000
75%	7489.250000	61.000000	42.000000	11.000000	17.000000	2.000000	1995.000000	2.000000	9.000000	3.000000	...	50.000000	20.000000	4.000000	1.000000	12.000000	28.000000	1.000000	6.000000	76.000000	10.000000
max	10002.000000	5326.000000	802.000000	44.000000	40.000000	6.000000	2015.000000	5.000000	123.000000	4.000000	...	636.000000	376.000000	147.000000	30.000000	151.000000	250.000000	2.000000	105.000000	218.000000	20.000000

8 rows × 275 columns

# housing[0].value_counts()
housing_train_num_str = num_pipeline.fit_transform(housing_train_num)  # housing_num为无文本数值的训练数据
housing_num_str = pd.DataFrame(housing_train_num_str, columns=housing_train_num.columns)
housing_num_str.describe()

	id	full_sq	life_sq	floor	max_floor	material	build_year	num_room	kitch_sq	state	...	cafe_count_5000_price_1500	cafe_count_5000_price_2500	cafe_count_5000_price_4000	cafe_count_5000_price_high	big_church_count_5000	church_count_5000	mosque_count_5000	leisure_count_5000	sport_count_5000	market_count_5000
count	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	...	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	8.000000e+03	8000.000000	8.000000e+03	8.000000e+03
mean	7.105427e-18	-1.243450e-17	-7.904788e-17	-7.815970e-17	-7.016610e-17	-2.406964e-16	-9.792167e-16	-1.749711e-16	2.153833e-17	4.325429e-16	...	-1.554312e-17	4.085621e-17	4.884981e-17	-1.509903e-17	3.463896e-17	7.993606e-18	-6.217249e-17	0.000000	2.664535e-18	2.131628e-17
std	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	...	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	1.000063e+00	1.000063	1.000063e+00	1.000063e+00
min	-1.718178e+00	-7.605148e-01	-1.606265e+00	-1.487177e+00	-8.967104e+00	-1.174785e-01	-4.453838e+01	-5.731910e+00	-4.289547e+00	-1.101076e+01	...	-5.162524e-01	-4.347192e-01	-3.746527e-01	-3.192044e-01	-5.202512e-01	-6.007020e-01	-8.014771e-01	-0.411373	-1.173476e+00	-1.259598e+00
25%	-8.662052e-01	-2.461927e-01	-5.913394e-01	-7.423666e-01	-2.241216e-03	-1.174785e-01	1.772092e-02	1.868945e-02	-4.667458e-03	1.196519e-01	...	-4.491267e-01	-4.063917e-01	-3.378820e-01	-3.192044e-01	-4.487427e-01	-4.693484e-01	-8.014771e-01	-0.411373	-9.066710e-01	-1.050027e+00
50%	2.564371e-04	-1.215086e-01	-1.563711e-01	-1.837587e-01	-2.241216e-03	-1.174785e-01	1.772092e-02	1.868945e-02	-4.667458e-03	1.196519e-01	...	-3.148752e-01	-3.072454e-01	-3.011112e-01	-3.192044e-01	-2.699714e-01	-2.942102e-01	-8.014771e-01	-0.310674	-1.284914e-01	-2.117448e-01
75%	8.649933e-01	1.122742e-01	1.819376e-01	5.610518e-01	-2.241216e-03	-1.174785e-01	1.772092e-02	1.868945e-02	-4.667458e-03	1.196519e-01	...	-9.671663e-02	-1.514442e-01	-2.275697e-01	-1.275620e-01	-9.120016e-02	-3.150298e-02	8.417359e-01	-0.109277	5.162861e-01	8.361079e-01
max	1.731800e+00	8.217002e+01	3.715424e+01	6.705738e+00	2.091577e+01	1.338579e+01	9.201230e-01	1.727049e+01	7.100191e+01	5.684856e+00	...	4.820243e+00	4.890852e+00	5.030650e+00	5.430067e+00	4.878641e+00	4.828581e+00	2.484949e+00	4.875309	3.673472e+00	2.931813e+00

8 rows × 275 columns

from sklearn.base import BaseEstimator, TransformerMixin
from sklearn. preprocessing import OneHotEncoder
from sklearn.pipeline import FeatureUnion

from sklearn.model_selection import GridSearchCV
from sklearn.linear_model import LinearRegression
from sklearn.svm import SVR
from sklearn.ensemble import RandomForestRegressor

import time

# scikit-learn暂时不支持DataFrame，因此我们需要分别选择数值型和枚举型
class DataFrameSelector(BaseEstimator, TransformerMixin):
    def __init__(self, attribute_names):
        self.attribute_names = attribute_names
    def fit(self, X, y=None):
        return self
    def transform(self, X):
        return X[self.attribute_names].values
    
num_pipeline = Pipeline([
    ('selector', DataFrameSelector(numeric_feature_names)),
    ('imputer', SimpleImputer(strategy="median")),
    ('std_scaler', StandardScaler())
])

cat_pipeline = Pipeline(steps=[
    ('selector', DataFrameSelector(text_feature_names)),
    ('imputer', SimpleImputer(strategy='most_frequent')),
    ('onehot', OneHotEncoder(handle_unknown='ignore'))
])

full_pipeline = FeatureUnion(transformer_list=[
    ("num_pipeline", num_pipeline),
    ("cat_pipeline", cat_pipeline)
])

full_pipeline_with_model = Pipeline(steps=[
    ("full_pipeline", full_pipeline),
    ("model", LinearRegression()),
])

param_grid = {
    "full_pipeline__num_pipeline__imputer__strategy": ["mean", "median", "most_frequent"],
    "full_pipeline__cat_pipeline__imputer__strategy": ["most_frequent", "constant"],
    "full_pipeline__cat_pipeline__onehot__handle_unknown": ['error', 'ignore'],
    "model": [LinearRegression(), SVR(), RandomForestRegressor()]
}


start = time.time()

# 创建 GridSearchCV 对象，进行参数搜索
grid_search = GridSearchCV(full_pipeline_with_model, param_grid=param_grid, cv=3)
grid_search.fit(housing_train, housing_train_labels)

end = time.time()
print("耗时：" + str(end-start))

# 打印最佳参数组合和模型得分
print("Best parameter combination: ", grid_search.best_params_)
print("Best model score: ", grid_search.best_score_)

F:\anaconda3\envs\TF2.1\lib\site-packages\sklearn\model_selection\_search.py:972: UserWarning: One or more of the test scores are non-finite: [        nan         nan         nan         nan         nan         nan
         nan         nan         nan -0.60319146 -0.02711282  0.52201065
 -0.61663247 -0.02711263  0.52200177 -0.65495276 -0.02711175  0.52994882
         nan         nan         nan         nan         nan         nan
         nan         nan         nan -0.60319146 -0.02711282  0.53401227
 -0.61663247 -0.02711263  0.53067201 -0.65495276 -0.02711175  0.52880227]
  category=UserWarning,


耗时：20075.620085954666
Best parameter combination:  {'full_pipeline__cat_pipeline__imputer__strategy': 'constant', 'full_pipeline__cat_pipeline__onehot__handle_unknown': 'ignore', 'full_pipeline__num_pipeline__imputer__strategy': 'mean', 'model': RandomForestRegressor()}

# 分析每个属性（列）的重要程度(把他们对应起来)
best_pipeline = grid_search.best_estimator_
best_model = best_pipeline.named_steps['model']
feature_importances = best_model.feature_importances_

feature_importances = feature_importances[:len(numeric_feature_names)]
print(sorted(zip(feature_importances, numeric_feature_names), reverse=True))

[(0.3558050288728259, 'full_sq'), (0.0369022366376152, 'cafe_count_2000'), (0.030728634527189217, 'office_sqm_5000'), (0.019937082781744157, 'cafe_count_3000_price_1500'), (0.015537904861632186, 'cafe_count_3000'), (0.013112393705737276, 'floor'), (0.01302595118237522, 'life_sq'), (0.011152233036002608, 'id'), (0.010880138446129202, 'cafe_count_2000_price_2500'), (0.008877243813506767, 'sport_count_3000'), (0.008849575375988353, 'sadovoe_km'), (0.00847971413204643, 'cafe_count_1000'), (0.007812304011587292, 'build_count_monolith'), (0.007124466078378038, 'theater_km'), (0.00598955666238343, 'green_part_1000'), (0.005293732058797867, 'cafe_count_5000_price_4000'), (0.004837345161800656, 'cafe_count_3000_price_2500'), (0.004817345418725336, 'kindergarten_km'), (0.00461492176396289, 'hospital_beds_raion'), (0.0045375430504188555, 'sport_count_2000'), (0.004470754833849411, 'preschool_quota'), (0.004425261691208156, 'public_healthcare_km'), (0.004284497060241596, 'big_road1_km'), (0.004248784991517519, 'zd_vokzaly_avto_km'), (0.004136713219380025, 'cafe_sum_1500_min_price_avg'), (0.0040865715928336014, 'cafe_count_1500'), (0.004079397582177327, 'additional_education_km'), (0.004052374972389635, 'big_road2_km'), (0.003988470054377696, 'ID_metro'), (0.0037952582929236, 'build_count_panel'), (0.003774838216804793, 'workplaces_km'), (0.0037429183086013025, 'swim_pool_km'), (0.0037373970684597656, 'power_transmission_line_km'), (0.0036912929593904283, 'metro_km_avto'), (0.0036524212108699876, 'green_part_2000'), (0.003637247698047196, 'public_transport_station_min_walk'), (0.0035343553043008886, 'market_shop_km'), (0.0035300284308685064, 'prom_part_3000'), (0.003493039231479483, 'green_zone_km'), (0.0034719231362119367, 'water_treatment_km'), (0.0034219894204335654, 'green_part_1500'), (0.003329615073905508, 'catering_km'), (0.003314536975526, 'public_transport_station_km'), (0.0033046821247699238, 'thermal_power_plant_km'), (0.003278634047964668, 'radiation_km'), (0.003228769536659676, 'church_synagogue_km'), (0.003166425810752949, 'big_market_km'), (0.003121693741607176, 'prom_part_1500'), (0.003092860614699901, 'detention_facility_km'), (0.0030660751025299326, 'cafe_count_1000_price_high'), (0.0030641894052107197, 'cafe_count_5000_price_2500'), (0.002982240570342268, 'industrial_km'), (0.0029721100255484362, 'oil_chemistry_km'), (0.0029372748333046216, 'ice_rink_km'), (0.0029037018536741265, 'hospice_morgue_km'), (0.0028979672492565438, 'basketball_km'), (0.0028716241247858025, 'preschool_km'), (0.0028520765357115613, 'shopping_centers_km'), (0.002827192165055018, 'cafe_sum_3000_min_price_avg'), (0.002767545941323787, 'fitness_km'), (0.0027459730591502386, 'cafe_sum_3000_max_price_avg'), (0.002737267343525625, 'cafe_count_5000_na_price'), (0.002666128595370512, 'nuclear_reactor_km'), (0.0026443891031494643, 'water_km'), (0.0025825863149320084, 'metro_min_avto'), (0.002551003232936204, 'big_church_km'), (0.0025421906600374853, 'ID_big_road1'), (0.0025005218540241976, 'exhibition_km'), (0.0024592455179383373, 'railroad_station_avto_km'), (0.002446947457164698, 'school_km'), (0.00243477689783016, 'trc_sqm_5000'), (0.002388639752287746, 'cafe_sum_2000_min_price_avg'), (0.002373529486919849, 'cemetery_km'), (0.0023705833591715717, 'railroad_km'), (0.0023601865395306318, 'cafe_count_1500_price_1000'), (0.0022960336723069886, 'cafe_avg_price_1500'), (0.0022944322375373204, 'green_part_3000'), (0.0022678600686821377, 'cafe_avg_price_2000'), (0.0022403779397926764, 'sport_count_5000'), (0.002222779912811367, 'ttk_km'), (0.0022130206446797107, 'university_km'), (0.002207947463485604, 'railroad_station_avto_min'), (0.0021962332587115666, 'ts_km'), (0.002169777156575615, 'metro_km_walk'), (0.002149595498746991, 'incineration_km'), (0.002123650071105262, 'cafe_avg_price_3000'), (0.002099515576433346, 'ID_railroad_station_walk'), (0.002099417314723335, 'green_part_5000'), (0.0020798017252723726, 'trc_sqm_1000'), (0.0020712663570771847, 'bus_terminal_avto_km'), (0.002063120390980617, 'cafe_sum_2000_max_price_avg'), (0.0020253531204588756, 'cafe_sum_500_min_price_avg'), (0.002010741544731312, 'stadium_km'), (0.0019952258027357635, 'museum_km'), (0.0019808672104866024, 'green_part_500'), (0.0019596999429266713, 'trc_sqm_3000'), (0.0019418970137407457, 'trc_count_2000'), (0.0018829571414917402, 'office_count_1000'), (0.001860785317751558, 'park_km'), (0.0018573122257889987, 'ID_railroad_station_avto'), (0.0018502586755312586, 'office_km'), (0.0018502048528180166, 'sport_count_1500'), (0.0017950315340824433, 'cafe_count_3000_na_price'), (0.0017408916406401887, 'cafe_count_1500_price_4000'), (0.0017066627368712514, 'cafe_count_1000_price_1500'), (0.0017012006774767565, 'cafe_count_5000_price_1000'), (0.0016788047956672809, 'cafe_sum_5000_min_price_avg'), (0.0016591314676513353, 'office_count_5000'), (0.0016215611970843834, 'metro_min_walk'), (0.0016027121916663456, 'trc_sqm_2000'), (0.0015896148083924218, 'prom_part_1000'), (0.001546125572060984, 'mosque_km'), (0.0015365685956578443, 'office_sqm_1500'), (0.00152326409219622, 'full_all'), (0.0015204454965024228, 'office_sqm_3000'), (0.0015057001257621492, 'trc_count_3000'), (0.001494633569870594, 'prom_part_5000'), (0.0014741025308991666, 'cafe_count_3000_price_500'), (0.0014712578299238576, 'trc_sqm_1500'), (0.0014707171791260405, 'prom_part_2000'), (0.0014706187046799005, 'trc_count_1500'), (0.0014375284468118282, 'mkad_km'), (0.0014370105029316513, 'cafe_count_500'), (0.0014207990144875333, 'indust_part'), (0.0014038140502102637, 'cafe_count_1000_price_1000'), (0.0013930160501839928, 'cafe_count_5000_price_500'), (0.0013868644433262223, 'railroad_station_walk_min'), (0.0013698389206839422, 'cafe_count_2000_price_1000'), (0.0013394166555226875, 'cafe_count_1000_price_500'), (0.0013386965146843976, 'office_sqm_2000'), (0.0013326936809903197, 'build_count_1946-1970'), (0.0013128109452357497, 'cafe_count_3000_price_4000'), (0.0012984499017887556, 'cafe_sum_5000_max_price_avg'), (0.0012800996165303555, 'cafe_count_1500_price_1500'), (0.0012761934657111557, 'build_count_after_1995'), (0.0012718569201853303, 'big_church_count_2000'), (0.00127058330811667, 'cafe_count_2000_price_500'), (0.0012638220027016128, 'ID_railroad_terminal'), (0.0012613567555310671, 'railroad_station_walk_km'), (0.0012434356062678146, 'green_zone_part'), (0.0012174455328585047, 'build_count_1971-1995'), (0.001214217771910042, 'sport_count_1000'), (0.0012140281465657175, 'cafe_avg_price_500'), (0.0012081753679400063, 'bulvar_ring_km'), (0.0012061047238333977, 'cafe_count_5000'), (0.0011839159920214273, 'healthcare_centers_raion'), (0.0011771248889729195, 'leisure_count_3000'), (0.001160775840178709, 'office_sqm_1000'), (0.0011551690564398636, 'cafe_count_1500_price_500'), (0.001132459081265372, 'cafe_avg_price_5000'), (0.001122449562282057, 'cafe_avg_price_1000'), (0.0011135964866840426, 'office_count_2000'), (0.0011111112115422372, 'cafe_sum_1500_max_price_avg'), (0.001069842749907959, 'school_quota'), (0.00106774249382398, 'cafe_count_2000_price_1500'), (0.0010602815927741205, 'cafe_sum_1000_max_price_avg'), (0.001044900091696954, 'cafe_sum_500_max_price_avg'), (0.001034667207937732, 'church_count_5000'), (0.0010301706062677523, 'cafe_count_5000_price_high'), (0.0010092727950096452, 'area_m'), (0.0009682554635303513, 'raion_build_count_with_material_info'), (0.0009587938492049883, '16_29_male'), (0.0009555688936662105, 'cafe_count_500_price_1500'), (0.0009535214961759879, 'ID_big_road2'), (0.000942206565069341, 'prom_part_500'), (0.0009311442185289545, 'cafe_count_500_price_2500'), (0.000925944901707187, 'cafe_count_3000_price_1000'), (0.000917036109269484, 'female_f'), (0.0009166333428757389, 'office_raion'), (0.0009037218668677034, 'trc_count_5000'), (0.0009017982067247552, 'build_count_brick'), (0.0009016525553024209, '16_29_female'), (0.0008981451459412267, 'additional_education_raion'), (0.0008892925799033124, 'sport_objects_raion'), (0.0008869222028668247, 'leisure_count_500'), (0.0008649458196326145, 'market_count_2000'), (0.0008606533432220762, 'cafe_count_1500_na_price'), (0.0008482599181997117, 'work_male'), (0.0008313163900783171, 'raion_build_count_with_builddate_info'), (0.0008176531565173631, 'church_count_2000'), (0.000798035936345229, 'leisure_count_1000'), (0.0007978367110980138, 'church_count_3000'), (0.0007886108904256832, 'trc_sqm_500'), (0.0007589258852283437, 'office_count_3000'), (0.0007532881423325608, '16_29_all'), (0.0007492697593058538, 'trc_count_1000'), (0.0007407201224415953, 'big_church_count_5000'), (0.0007358415153196833, 'office_count_1500'), (0.0007343517710255492, 'shopping_centers_raion'), (0.0007197409542981459, 'cafe_count_2000_na_price'), (0.0007185895688121112, 'school_education_centers_raion'), (0.0007028017150241994, 'cafe_count_500_price_500'), (0.0007016613633910146, 'cafe_sum_1000_min_price_avg'), (0.0006943690591683538, 'market_count_5000'), (0.0006844796452933984, 'office_sqm_500'), (0.0006764053215622059, 'male_f'), (0.0006759856898131497, 'cafe_count_2000_price_4000'), (0.0006737379724652002, 'church_count_500'), (0.0006643541654161514, 'leisure_count_2000'), (0.0006594348467692361, 'cafe_count_5000_price_1500'), (0.0006535942774188417, 'cafe_count_1000_price_2500'), (0.0006450610078269494, 'cafe_count_1000_na_price'), (0.0006346299560631147, 'big_church_count_3000'), (0.0006268199748514777, 'market_count_1000'), (0.0006136616922540099, 'cafe_count_1500_price_2500'), (0.0006101020772724982, 'build_count_1921-1945'), (0.000605333476588235, 'build_count_slag'), (0.0006025808857398215, 'sport_count_500'), (0.0005913398638573828, 'leisure_count_1500'), (0.0005893053464770819, '0_13_all'), (0.0005726863992438555, 'office_count_500'), (0.0005686382393212504, 'cafe_count_500_price_1000'), (0.0005335367189856385, 'preschool_education_centers_raion'), (0.0005281848859666116, 'build_count_block'), (0.0005053507999611831, 'church_count_1500'), (0.000503907065993229, 'young_female'), (0.0004944574646751173, 'cafe_count_500_price_4000'), (0.0004940955145651672, 'build_count_before_1920'), (0.0004563872071786454, '7_14_female'), (0.0004467097825214889, 'ID_bus_terminal'), (0.00044440004204567495, 'kremlin_km'), (0.0004245073138422882, 'market_count_3000'), (0.00041927115333189414, 'church_count_1000'), (0.00039153288360794847, 'ekder_male'), (0.0003894559280018075, 'leisure_count_5000'), (0.00035520364499142795, 'cafe_count_500_na_price'), (0.0003426960631525808, 'cafe_count_3000_price_high'), (0.0003358805974856705, 'ekder_female'), (0.0003354447960831321, 'market_count_1500'), (0.0003161403348918282, 'work_female'), (0.00031513222155404733, 'trc_count_500'), (0.0002789002146380332, 'big_church_count_1000'), (0.00026355257451541296, 'cafe_count_1000_price_4000'), (0.0002592117928358345, 'work_all'), (0.000246114632595913, 'ekder_all'), (0.0002355341717062321, 'mosque_count_5000'), (0.00023529864017649287, 'raion_popul'), (0.0002271843949564929, 'children_school'), (0.0002250069306485336, 'kitch_sq'), (0.00022431385558648625, 'children_preschool'), (0.00022412795110272134, 'max_floor'), (0.00022004001831562965, '0_13_male'), (0.0002177633191441963, '7_14_all'), (0.000211450005150454, 'school_education_centers_top_20_raion'), (0.00021129814290958885, 'build_count_wood'), (0.00020721481519426333, 'build_year'), (0.0002011649065965163, 'big_church_count_1500'), (0.00019640473122088542, '0_6_all'), (0.0001959025906314739, 'build_count_frame'), (0.00019056850712193235, 'big_church_count_500'), (0.00018503977876596978, '0_6_male'), (0.00018348735544191697, '0_17_all'), (0.0001816468582567304, '7_14_male'), (0.0001764278041528311, 'university_top_20_raion'), (0.00017137088658084054, 'cafe_count_2000_price_high'), (0.0001689115793605786, '0_6_female'), (0.00016185056709061189, 'market_count_500'), (0.00015899427406858044, 'num_room'), (0.0001502293154301887, '0_17_male'), (0.00014287804669518018, 'young_all'), (0.00013951399639071044, 'state'), (0.00013783866519237893, 'material'), (0.00013297476626663562, '0_17_female'), (0.00011858523670534393, 'young_male'), (0.00010142842124175284, 'culture_objects_top_25_raion'), (9.525954968205083e-05, '0_13_female'), (9.499848473560952e-05, 'mosque_count_1000'), (9.488543927847641e-05, 'cafe_count_1500_price_high'), (7.633784366627806e-05, 'mosque_count_3000'), (5.9661958682616646e-05, 'build_count_foam'), (4.7001210819952146e-05, 'build_count_mix'), (3.5218669751835894e-05, 'mosque_count_1500'), (3.5050631496159006e-05, 'mosque_count_500'), (2.257946417881287e-05, 'mosque_count_2000'), (1.2116849876895503e-05, 'cafe_count_500_price_high')]

# 训练集上的误差
from sklearn.metrics import mean_squared_error

final_model = grid_search.best_estimator_

housing_train_prepare = full_pipeline.fit_transform(housing_train)

Train_predictions = final_model.predict(housing_train)

Train_mse = mean_squared_error(housing_train_labels, Train_predictions)
Train_rmse = np.sqrt(Train_mse)

print(Train_rmse)
# 房价前25%：4.233045e+06

1132793.1642235247

# 测试集上的误差
X_test = strat_test_set.drop("price_doc", axis=1)
Y_test = strat_test_set["price_doc"].copy()

final_predictions = final_model.predict(X_test)

final_mse = mean_squared_error(Y_test, final_predictions)
final_rmse = np.sqrt(final_mse)

print(final_rmse)

2562534.086991416

# rmse是泛化误差，估计的精确度可以使用scipy计算置信区间
from scipy import stats

confidence = 0.95
squared_errors = (final_predictions-Y_test)**2
num = np.sqrt(stats.t.interval(confidence, len(squared_errors)-1,
                             loc=squared_errors.mean(),
                             scale = stats.sem(squared_errors)))
print(num)

[2206930.88712145 2874476.98781104]

如果跑完了，就从In[21]开始：housing_num->housing.describe()->num_pipline.fit_transform(housing_num) ->housing.describe()
看看有没有填成功。然后处理文本，看看增加了多少个one-hot维度
from sklearn.pipeline import Pipeline
from sklearn.impute import SimpleImputer 
from sklearn.preprocessing import StandardScaler, OneHotEncoder
from sklearn.compose import ColumnTransformer
from sklearn.model_selection import GridSearchCV
from sklearn.linear_model import LinearRegression
from sklearn.svm import SVR
from sklearn.ensemble import RandomForestRegressor

# 定义数据准备和模型训练的 Pipeline
numeric_transformer = Pipeline(steps=[
    ('imputer', SimpleImputer()),
    ('scaler', StandardScaler())
])

categorical_transformer = Pipeline(steps=[
    ('imputer', SimpleImputer(strategy='most_frequent')),
    ('onehot', OneHotEncoder(handle_unknown='ignore'))
])

preprocessor = ColumnTransformer(
    transformers=[
        ('num', numeric_transformer, numeric_feature_names),
        ('cat', categorical_transformer, text_feature_names)
    ])

pipeline = Pipeline(steps=[('preprocessor', preprocessor),
                            ('regressor', LinearRegression())])

# 定义参数网格用于搜索最佳参数组合
param_grid = {
    'preprocessor__num__imputer__strategy': ['mean', 'median', 'most_frequent', None],
    'preprocessor__cat__imputer__strategy': ['mean', 'median', 'most_frequent', None],
    'preprocessor__cat__onehot__handle_unknown': ['error', 'ignore'],
    'regressor': [LinearRegression(), SVR(), RandomForestRegressor()],
}

# print("hello")

import time

start = time.time()

# 创建 GridSearchCV 对象，进行参数搜索
grid_search = GridSearchCV(pipeline, param_grid=param_grid, cv=3)
grid_search.fit(housing_train, housing_train_labels)

end = time.time()
print("耗时：" + str(end-start))
# 打印最佳参数组合和模型得分
print("Best parameter combination: ", grid_search.best_params_)
print("Best model score: ", grid_search.best_score_)

# 耗时：7641.132771015167
# Best parameter combination:  {'preprocessor__cat__imputer__strategy': 'most_frequent', 'preprocessor__cat__onehot__handle_unknown': 'ignore', 'preprocessor__num__imputer__strategy': 'median', 'regressor': RandomForestRegressor()}
# Best model score:  0.5281557991343143

F:\anaconda3\envs\TF2.1\lib\site-packages\sklearn\model_selection\_search.py:972: UserWarning: One or more of the test scores are non-finite: [        nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
 -0.60319146 -0.02711282  0.52238756 -0.61663247 -0.02711263  0.5281558
 -0.65495276 -0.02711175  0.52238873         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan
         nan         nan         nan         nan         nan         nan]
  category=UserWarning,


耗时：7641.132771015167
Best parameter combination:  {'preprocessor__cat__imputer__strategy': 'most_frequent', 'preprocessor__cat__onehot__handle_unknown': 'ignore', 'preprocessor__num__imputer__strategy': 'median', 'regressor': RandomForestRegressor()}
Best model score:  0.5281557991343143