Data Every Day: 不動産データセット

tldr

KggleのReal Estate DataSetをPredicting Real Estate Value With Random Forests - Data Every Day #086に沿ってやっていきます。

実行環境はGoogle Colaboratorです。

インポート

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

import sklearn.preprocessing as sp
from sklearn.model_selection import train_test_split
import sklearn.linear_model as slm
from sklearn.ensemble import RandomForestRegressor

データのダウンロード

Google Driveをマウントします。

from google.colab import drive
drive.mount('/content/drive')

Mounted at /content/drive

KaggleのAPIクライアントを初期化し、認証します。
認証情報はGoogle Drive内（/content/drive/My Drive/Colab Notebooks/Kaggle）にkaggle.jsonとして置いてあります。

import os
kaggle_path = "/content/drive/My Drive/Colab Notebooks/Kaggle"
os.environ['KAGGLE_CONFIG_DIR'] = kaggle_path

from kaggle.api.kaggle_api_extended import KaggleApi
api = KaggleApi()
api.authenticate() 

Kaggle APIを使ってデータをダウンロードします。

dataset_id = 'arslanali4343/real-estate-dataset'
dataset = api.dataset_list_files(dataset_id)
file_name = dataset.files[0].name
file_path = os.path.join(api.get_default_download_dir(), file_name)
file_path

Warning: Looks like you're using an outdated API Version, please consider updating (server 1.5.10 / client 1.5.9)





'/content/data.csv'

api.dataset_download_file(dataset_id, file_name, force=True, quiet=False)

100%|██████████| 34.6k/34.6k [00:00<00:00, 5.74MB/s]

Downloading data.csv to /content









True

データの読み込み

Padasを使ってダウンロードしてきたCSVファイルを読み込みます。

data = pd.read_csv(file_path)

data

	CRIM	ZN	INDUS	CHAS	NOX	RM	AGE	DIS	RAD	TAX	PTRATIO	B	LSTAT	MEDV
0	0.00632	18.0	2.31	0	0.538	6.575	65.2	4.0900	1	296	15.3	396.90	4.98	24.0
1	0.02731	0.0	7.07	0	0.469	6.421	78.9	4.9671	2	242	17.8	396.90	9.14	21.6
2	0.02729	0.0	7.07	0	0.469	7.185	61.1	4.9671	2	242	17.8	392.83	4.03	34.7
3	0.03237	0.0	2.18	0	0.458	6.998	45.8	6.0622	3	222	18.7	394.63	2.94	33.4
4	0.06905	0.0	2.18	0	0.458	7.147	54.2	6.0622	3	222	18.7	396.90	5.33	36.2
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
506	0.98765	0.0	12.50	0	0.561	6.980	89.0	2.0980	3	320	23.0	396.00	12.00	12.0
507	0.23456	0.0	12.50	0	0.561	6.980	76.0	2.6540	3	320	23.0	343.00	25.00	32.0
508	0.44433	0.0	12.50	0	0.561	6.123	98.0	2.9870	3	320	23.0	343.00	21.00	54.0
509	0.77763	0.0	12.70	0	0.561	6.222	34.0	2.5430	3	329	23.0	343.00	76.00	67.0
510	0.65432	0.0	12.80	0	0.561	6.760	67.0	2.9870	3	345	23.0	321.00	45.00	24.0

511 rows × 14 columns

下準備

data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 511 entries, 0 to 510
Data columns (total 14 columns):
 #   Column   Non-Null Count  Dtype  
---  ------   --------------  -----  
 0   CRIM     511 non-null    float64
 1   ZN       511 non-null    float64
 2   INDUS    511 non-null    float64
 3   CHAS     511 non-null    int64  
 4   NOX      511 non-null    float64
 5   RM       506 non-null    float64
 6   AGE      511 non-null    float64
 7   DIS      511 non-null    float64
 8   RAD      511 non-null    int64  
 9   TAX      511 non-null    int64  
 10  PTRATIO  511 non-null    float64
 11  B        511 non-null    float64
 12  LSTAT    511 non-null    float64
 13  MEDV     511 non-null    float64
dtypes: float64(11), int64(3)
memory usage: 56.0 KB

欠損値の処理

data['RM'] = data['RM'].fillna(data['RM'].mean())

data.isna().sum().sum()

0

X, Yデータの分割

y = data['MEDV']
X = data.drop('MEDV', axis=1)

スケーリング

scaler = sp.StandardScaler()
X = scaler.fit_transform(X)

トレーニング、テストデータの分割

X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.7)

トレーニング

no_bootstrap_model = RandomForestRegressor(n_estimators=100, criterion='mse', bootstrap=False)
no_bootstrap_model.fit(X_train, y_train)

RandomForestRegressor(bootstrap=False, ccp_alpha=0.0, criterion='mse',
                      max_depth=None, max_features='auto', max_leaf_nodes=None,
                      max_samples=None, min_impurity_decrease=0.0,
                      min_impurity_split=None, min_samples_leaf=1,
                      min_samples_split=2, min_weight_fraction_leaf=0.0,
                      n_estimators=100, n_jobs=None, oob_score=False,
                      random_state=None, verbose=0, warm_start=False)

bootstrap_model = RandomForestRegressor(n_estimators=100, criterion='mse', bootstrap=True)
bootstrap_model.fit(X_train, y_train)

RandomForestRegressor(bootstrap=True, ccp_alpha=0.0, criterion='mse',
                      max_depth=None, max_features='auto', max_leaf_nodes=None,
                      max_samples=None, min_impurity_decrease=0.0,
                      min_impurity_split=None, min_samples_leaf=1,
                      min_samples_split=2, min_weight_fraction_leaf=0.0,
                      n_estimators=100, n_jobs=None, oob_score=False,
                      random_state=None, verbose=0, warm_start=False)

結果

print('R^2 without bootstrapping:', no_bootstrap_model.score(X_test, y_test))
print('   R^2 with bootstrapping:', bootstrap_model.score(X_test, y_test))

R^2 without bootstrapping: 0.46518356517728865
   R^2 with bootstrapping: 0.759932629153558

ブートストラップサンプリングを使ったモデルがかなり高い精度をだしました。