Data Science/파이썬 머신러닝 완벽 가이드
[Dacon] 보스턴 집값 예측
얆생
2023. 4. 20. 22:01
https://dacon.io/competitions/open/236065/overview/description
보스턴 집값 예측 경진대회 - DACON
분석시각화 대회 코드 공유 게시물은 내용 확인 후 좋아요(투표) 가능합니다.
dacon.io
데이콘에서 교육용 연습대회에 참가해보았다.
- CRIM: 도시별 1인당 범죄율
- ZN: 25,000 피트를 초과하는 주거용 토지의 비율
- NDUS: 비상업 면적의 비율
- CHAS: 찰스강에 대한 더미 변수(강의 경계에 위치한 경우는 1, 아니면 0)
- NOX: 일산화질소 농도
- RM: 주택당 평균 방의 개수
- AGE: 1940년 이전에 건축된 자가주택의 비율
- DIS: 5개의 보스턴 고용 센터와의 거리
- RAD: 고속도로 접근성 지수
- TAX: 10,000달러당 재산세율
- PTRATIO: 도시별 교사와 학생 수 비율
- B: 마을의 흑인 거주 비율
- LSTAT: 하위 계층의 비율
- MEDV: 본인 소유의 주택 가격(중앙값, 단위: 천달러)
In [1]:
import pandas as pd
import seaborn as sns
import statsmodels.api as sm
import statsmodels.formula.api as smf
from statsmodels.formula.api import ols, glm
house = pd.read_csv('./Boston Housing 데이터/train.csv')
house
Out[1]:
| ID | CRIM | ZN | INDUS | CHAS | NOX | RM | AGE | DIS | RAD | TAX | PTRATIO | B | LSTAT | MEDV | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | TRAIN_000 | 0.04819 | 80.0 | 3.64 | 0.0 | 0.392 | 6.108 | 32.0 | 9.2203 | 1.0 | 315.0 | 16.4 | 392.89 | 6.57 | 21.9 |
| 1 | TRAIN_001 | 1.42502 | 0.0 | 19.58 | 0.0 | 0.871 | 6.510 | 100.0 | 1.7659 | 5.0 | 403.0 | 14.7 | 364.31 | 7.39 | 23.3 |
| 2 | TRAIN_002 | 0.01778 | 95.0 | 1.47 | 0.0 | 0.403 | 7.135 | 13.9 | 7.6534 | 3.0 | 402.0 | 17.0 | 384.30 | 4.45 | 32.9 |
| 3 | TRAIN_003 | 9.51363 | 0.0 | 18.10 | 0.0 | 0.713 | 6.728 | 94.1 | 2.4961 | 24.0 | 666.0 | 20.2 | 6.68 | 18.71 | 14.9 |
| 4 | TRAIN_004 | 1.65660 | 0.0 | 19.58 | 0.0 | 0.871 | 6.122 | 97.3 | 1.6180 | 5.0 | 403.0 | 14.7 | 372.80 | 14.10 | 21.5 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 450 | TRAIN_450 | 0.33983 | 22.0 | 5.86 | 0.0 | 0.431 | 6.108 | 34.9 | 8.0555 | 7.0 | 330.0 | 19.1 | 390.18 | 9.16 | 24.3 |
| 451 | TRAIN_451 | 3.83684 | 0.0 | 18.10 | 0.0 | 0.770 | 6.251 | 91.1 | 2.2955 | 24.0 | 666.0 | 20.2 | 350.65 | 14.19 | 19.9 |
| 452 | TRAIN_452 | 0.20746 | 0.0 | 27.74 | 0.0 | 0.609 | 5.093 | 98.0 | 1.8226 | 4.0 | 711.0 | 20.1 | 318.43 | 29.68 | 8.1 |
| 453 | TRAIN_453 | 0.09299 | 0.0 | 25.65 | 0.0 | 0.581 | 5.961 | 92.9 | 2.0869 | 2.0 | 188.0 | 19.1 | 378.09 | 17.93 | 20.5 |
| 454 | TRAIN_454 | 25.94060 | 0.0 | 18.10 | 0.0 | 0.679 | 5.304 | 89.1 | 1.6475 | 24.0 | 666.0 | 20.2 | 127.36 | 26.64 | 10.4 |
455 rows × 15 columns
In [2]:
house.describe()
Out[2]:
| CRIM | ZN | INDUS | CHAS | NOX | RM | AGE | DIS | RAD | TAX | PTRATIO | B | LSTAT | MEDV | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| count | 455.000000 | 455.000000 | 455.000000 | 455.000000 | 455.000000 | 455.000000 | 455.000000 | 455.000000 | 455.000000 | 455.000000 | 455.000000 | 455.000000 | 455.000000 | 455.000000 |
| mean | 3.659998 | 11.165934 | 11.092747 | 0.065934 | 0.555160 | 6.285352 | 68.320440 | 3.788439 | 9.551648 | 407.892308 | 18.453846 | 356.285714 | 12.604967 | 22.574725 |
| std | 8.888075 | 22.953757 | 6.821697 | 0.248440 | 0.116524 | 0.699346 | 27.893807 | 2.082031 | 8.718624 | 169.446805 | 2.149137 | 92.664335 | 7.090388 | 9.120759 |
| min | 0.006320 | 0.000000 | 0.460000 | 0.000000 | 0.389000 | 3.561000 | 2.900000 | 1.129600 | 1.000000 | 187.000000 | 12.600000 | 0.320000 | 1.730000 | 5.000000 |
| 25% | 0.079230 | 0.000000 | 5.160000 | 0.000000 | 0.449000 | 5.881000 | 44.700000 | 2.100350 | 4.000000 | 278.000000 | 17.400000 | 375.990000 | 6.970000 | 17.100000 |
| 50% | 0.245220 | 0.000000 | 9.690000 | 0.000000 | 0.538000 | 6.208000 | 76.700000 | 3.199200 | 5.000000 | 330.000000 | 19.000000 | 391.700000 | 11.280000 | 21.200000 |
| 75% | 3.685665 | 12.500000 | 18.100000 | 0.000000 | 0.624000 | 6.627000 | 93.850000 | 5.164950 | 24.000000 | 666.000000 | 20.200000 | 396.375000 | 16.695000 | 25.100000 |
| max | 88.976200 | 100.000000 | 27.740000 | 1.000000 | 0.871000 | 8.780000 | 100.000000 | 12.126500 | 24.000000 | 711.000000 | 22.000000 | 396.900000 | 36.980000 | 50.000000 |
In [3]:
print(house.groupby('CHAS')[['MEDV']].describe().unstack('CHAS'))
CHAS
MEDV count 0.0 425.000000
1.0 30.000000
mean 0.0 22.134824
1.0 28.806667
std 0.0 8.762722
1.0 11.690165
min 0.0 5.000000
1.0 13.400000
25% 0.0 16.700000
1.0 21.700000
50% 0.0 21.000000
1.0 23.850000
75% 0.0 24.800000
1.0 33.175000
max 0.0 50.000000
1.0 50.000000
dtype: float64
In [4]:
print(house.RAD.value_counts())
print(house.groupby('RAD')[['MEDV']].describe().unstack('RAD'))
24.0 119
5.0 106
4.0 96
3.0 36
6.0 25
2.0 21
8.0 20
1.0 18
7.0 14
Name: RAD, dtype: int64
RAD
MEDV count 1.0 18.0
2.0 21.0
3.0 36.0
4.0 96.0
5.0 106.0
...
max 5.0 50.0
6.0 24.8
7.0 42.8
8.0 50.0
24.0 50.0
Length: 72, dtype: float64
In [5]:
#상관관계 분석
house.corr()
C:\Users\Administrator\AppData\Local\Temp\ipykernel_28408\3222390172.py:2: FutureWarning: The default value of numeric_only in DataFrame.corr is deprecated. In a future version, it will default to False. Select only valid columns or specify the value of numeric_only to silence this warning.
house.corr()
Out[5]:
| CRIM | ZN | INDUS | CHAS | NOX | RM | AGE | DIS | RAD | TAX | PTRATIO | B | LSTAT | MEDV | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CRIM | 1.000000 | -0.196246 | 0.404043 | -0.053530 | 0.410551 | -0.212673 | 0.349981 | -0.375011 | 0.614488 | 0.570890 | 0.287082 | -0.384054 | 0.444276 | -0.386511 |
| ZN | -0.196246 | 1.000000 | -0.525693 | -0.025098 | -0.505469 | 0.317844 | -0.556758 | 0.647406 | -0.306968 | -0.316827 | -0.388047 | 0.181655 | -0.416454 | 0.376084 |
| INDUS | 0.404043 | -0.525693 | 1.000000 | 0.057221 | 0.759181 | -0.391466 | 0.638347 | -0.702260 | 0.603900 | 0.730394 | 0.373639 | -0.369784 | 0.608342 | -0.483034 |
| CHAS | -0.053530 | -0.025098 | 0.057221 | 1.000000 | 0.081192 | 0.121848 | 0.069445 | -0.089970 | -0.006660 | -0.033841 | -0.118460 | 0.043095 | -0.064032 | 0.181734 |
| NOX | 0.410551 | -0.505469 | 0.759181 | 0.081192 | 1.000000 | -0.306085 | 0.730537 | -0.762721 | 0.609330 | 0.666784 | 0.175569 | -0.385323 | 0.597259 | -0.433293 |
| RM | -0.212673 | 0.317844 | -0.391466 | 0.121848 | -0.306085 | 1.000000 | -0.245285 | 0.210719 | -0.214426 | -0.299118 | -0.354118 | 0.131836 | -0.611582 | 0.697426 |
| AGE | 0.349981 | -0.556758 | 0.638347 | 0.069445 | 0.730537 | -0.245285 | 1.000000 | -0.741336 | 0.462316 | 0.517079 | 0.250436 | -0.281139 | 0.603303 | -0.384123 |
| DIS | -0.375011 | 0.647406 | -0.702260 | -0.089970 | -0.762721 | 0.210719 | -0.741336 | 1.000000 | -0.496010 | -0.539764 | -0.227028 | 0.302359 | -0.498786 | 0.260402 |
| RAD | 0.614488 | -0.306968 | 0.603900 | -0.006660 | 0.609330 | -0.214426 | 0.462316 | -0.496010 | 1.000000 | 0.908535 | 0.467822 | -0.454256 | 0.486773 | -0.383792 |
| TAX | 0.570890 | -0.316827 | 0.730394 | -0.033841 | 0.666784 | -0.299118 | 0.517079 | -0.539764 | 0.908535 | 1.000000 | 0.461456 | -0.450327 | 0.547234 | -0.470998 |
| PTRATIO | 0.287082 | -0.388047 | 0.373639 | -0.118460 | 0.175569 | -0.354118 | 0.250436 | -0.227028 | 0.467822 | 0.461456 | 1.000000 | -0.183874 | 0.362355 | -0.508638 |
| B | -0.384054 | 0.181655 | -0.369784 | 0.043095 | -0.385323 | 0.131836 | -0.281139 | 0.302359 | -0.454256 | -0.450327 | -0.183874 | 1.000000 | -0.366402 | 0.332199 |
| LSTAT | 0.444276 | -0.416454 | 0.608342 | -0.064032 | 0.597259 | -0.611582 | 0.603303 | -0.498786 | 0.486773 | 0.547234 | 0.362355 | -0.366402 | 1.000000 | -0.738124 |
| MEDV | -0.386511 | 0.376084 | -0.483034 | 0.181734 | -0.433293 | 0.697426 | -0.384123 | 0.260402 | -0.383792 | -0.470998 | -0.508638 | 0.332199 | -0.738124 | 1.000000 |
In [15]:
import seaborn as sns
sns.heatmap(data=house.corr(), cmap='Blues')
C:\Users\Administrator\AppData\Local\Temp\ipykernel_28408\3515544278.py:2: FutureWarning: The default value of numeric_only in DataFrame.corr is deprecated. In a future version, it will default to False. Select only valid columns or specify the value of numeric_only to silence this warning.
sns.heatmap(data=house.corr(), cmap='Blues')
Out[15]:
<AxesSubplot: >
- MEDV랑 RM: 강한 양의 상관관계 >> 주택당 평균 방의 개수가 많을 수록 집값이 비싸다
- MEDV랑 LSTAT: 강한 음의 상관관계 >> 하위계층의 비율이 높을수록 집값이 싸다.
In [6]:
#회귀분석
my_formula = 'MEDV ~ CRIM + ZN + INDUS + NOX + RM + AGE + DIS + RAD + TAX + PTRATIO + B + LSTAT'
lm = ols(my_formula, data=house).fit()
print(lm.summary())
print('\nQuantities you can extract from the result: \n%s' % dir(lm))
print('\nCoefficients: \n%s' % lm.params)
print('\nCoefficient Std Errors: \n%s' % lm.bse)
print('\nAdj. R-sqaured: \n%.2f' % lm.rsquared_adj)
print('\nF-statistic: %.1f P-value: %.2f' % (lm.fvalue, lm.f_pvalue))
print('\nNumber of obs: %d Number of fitted values: %s' % (lm.nobs, len(lm.fittedvalues)))
OLS Regression Results
==============================================================================
Dep. Variable: MEDV R-squared: 0.738
Model: OLS Adj. R-squared: 0.731
Method: Least Squares F-statistic: 103.7
Date: Thu, 20 Apr 2023 Prob (F-statistic): 3.10e-120
Time: 19:04:03 Log-Likelihood: -1346.3
No. Observations: 455 AIC: 2719.
Df Residuals: 442 BIC: 2772.
Df Model: 12
Covariance Type: nonrobust
==============================================================================
coef std err t P>|t| [0.025 0.975]
------------------------------------------------------------------------------
Intercept 36.6451 5.365 6.830 0.000 26.101 47.189
CRIM -0.1145 0.033 -3.477 0.001 -0.179 -0.050
ZN 0.0462 0.014 3.223 0.001 0.018 0.074
INDUS 0.0682 0.064 1.059 0.290 -0.058 0.195
NOX -17.5083 3.957 -4.425 0.000 -25.285 -9.732
RM 3.9141 0.438 8.944 0.000 3.054 4.774
AGE 0.0030 0.014 0.220 0.826 -0.024 0.030
DIS -1.3816 0.206 -6.693 0.000 -1.787 -0.976
RAD 0.3225 0.068 4.710 0.000 0.188 0.457
TAX -0.0131 0.004 -3.392 0.001 -0.021 -0.006
PTRATIO -1.0375 0.138 -7.508 0.000 -1.309 -0.766
B 0.0093 0.003 3.319 0.001 0.004 0.015
LSTAT -0.5278 0.053 -9.910 0.000 -0.632 -0.423
==============================================================================
Omnibus: 187.867 Durbin-Watson: 1.973
Prob(Omnibus): 0.000 Jarque-Bera (JB): 1007.188
Skew: 1.724 Prob(JB): 1.96e-219
Kurtosis: 9.421 Cond. No. 1.50e+04
==============================================================================
Notes:
[1] Standard Errors assume that the covariance matrix of the errors is correctly specified.
[2] The condition number is large, 1.5e+04. This might indicate that there are
strong multicollinearity or other numerical problems.
Quantities you can extract from the result:
['HC0_se', 'HC1_se', 'HC2_se', 'HC3_se', '_HCCM', '__class__', '__delattr__', '__dict__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__lt__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', '_abat_diagonal', '_cache', '_data_attr', '_data_in_cache', '_get_robustcov_results', '_is_nested', '_use_t', '_wexog_singular_values', 'aic', 'bic', 'bse', 'centered_tss', 'compare_f_test', 'compare_lm_test', 'compare_lr_test', 'condition_number', 'conf_int', 'conf_int_el', 'cov_HC0', 'cov_HC1', 'cov_HC2', 'cov_HC3', 'cov_kwds', 'cov_params', 'cov_type', 'df_model', 'df_resid', 'diagn', 'eigenvals', 'el_test', 'ess', 'f_pvalue', 'f_test', 'fittedvalues', 'fvalue', 'get_influence', 'get_prediction', 'get_robustcov_results', 'info_criteria', 'initialize', 'k_constant', 'llf', 'load', 'model', 'mse_model', 'mse_resid', 'mse_total', 'nobs', 'normalized_cov_params', 'outlier_test', 'params', 'predict', 'pvalues', 'remove_data', 'resid', 'resid_pearson', 'rsquared', 'rsquared_adj', 'save', 'scale', 'ssr', 'summary', 'summary2', 't_test', 't_test_pairwise', 'tvalues', 'uncentered_tss', 'use_t', 'wald_test', 'wald_test_terms', 'wresid']
Coefficients:
Intercept 36.645120
CRIM -0.114515
ZN 0.046238
INDUS 0.068209
NOX -17.508327
RM 3.914144
AGE 0.003044
DIS -1.381636
RAD 0.322528
TAX -0.013146
PTRATIO -1.037455
B 0.009260
LSTAT -0.527760
dtype: float64
Coefficient Std Errors:
Intercept 5.364928
CRIM 0.032940
ZN 0.014348
INDUS 0.064428
NOX 3.956786
RM 0.437609
AGE 0.013845
DIS 0.206436
RAD 0.068470
TAX 0.003876
PTRATIO 0.138185
B 0.002790
LSTAT 0.053254
dtype: float64
Adj. R-sqaured:
0.73
F-statistic: 103.7 P-value: 0.00
Number of obs: 455 Number of fitted values: 455
- RM(주택당 평균 방의 개수)이 1단위 증가하면 MEDV(집값)이 평균 3.914 증가함
- LSTAT(하위계층 비율)이 1단위 증가하면 MEDV(집값)이 평균 0.527 하락함
In [7]:
#예측하기
test_data = pd.read_csv('./Boston Housing 데이터/test.csv')
predicted = lm.predict(test_data)
predicted_rounded = [round(score, 2) for score in predicted]
print(predicted)
In [9]:
import pandas as pd
submit = pd.read_csv('./Boston Housing 데이터/sample_submission.csv')
In [11]:
submit['MEDV'] = predicted
submit.head()
Out[11]:
| ID | MEDV | |
|---|---|---|
| 0 | TEST_000 | 22.724850 |
| 1 | TEST_001 | 32.277223 |
| 2 | TEST_002 | 19.828197 |
| 3 | TEST_003 | 24.575220 |
| 4 | TEST_004 | 26.172434 |
In [13]:
submit.to_csv('./submit.csv', index=False)
출력값을 csv파일로 저장하여 제출한 결과, 리더보드에서 36등을 하였다.
