[빅.분.기] 작업형2유형 - Logistic회귀

Machine Learning #1 : 분류 문제 - Logistic 회귀¶

자료 출처 : Datacampus "빅데이터 분석기사 자격증 과정 실기" 책 예제 : https://www.datacampus.co.kr/board/read.jsp?id=98394&code=notice

data/ library import¶

In [135]:

import warnings

In [136]:

warnings.filterwarnings('ignore')

In [3]:

import pandas as pd

data= pd.read_csv('breast-cancer-wisconsin.csv')

data 확인¶

In [4]:

data.head()

Out[4]:

	code	Clump_Thickness	Cell_Size	Cell_Shape	Marginal_Adhesion	Single_Epithelial_Cell_Size	Bare_Nuclei	Bland_Chromatin	Normal_Nucleoli	Mitoses	Class
0	1000025	5	1	1	1	2	1	3	1	1	0
1	1002945	5	4	4	5	7	10	3	2	1	0
2	1015425	3	1	1	1	2	2	3	1	1	0
3	1016277	6	8	8	1	3	4	3	7	1	0
4	1017023	4	1	1	3	2	1	3	1	1	0

In [5]:

data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 683 entries, 0 to 682
Data columns (total 11 columns):
 #   Column                       Non-Null Count  Dtype
---  ------                       --------------  -----
 0   code                         683 non-null    int64
 1   Clump_Thickness              683 non-null    int64
 2   Cell_Size                    683 non-null    int64
 3   Cell_Shape                   683 non-null    int64
 4   Marginal_Adhesion            683 non-null    int64
 5   Single_Epithelial_Cell_Size  683 non-null    int64
 6   Bare_Nuclei                  683 non-null    int64
 7   Bland_Chromatin              683 non-null    int64
 8   Normal_Nucleoli              683 non-null    int64
 9   Mitoses                      683 non-null    int64
 10  Class                        683 non-null    int64
dtypes: int64(11)
memory usage: 58.8 KB

In [7]:

data.describe().transpose()

Out[7]:

	count	mean	std	min	25%	50%	75%	max
code	683.0	1.076720e+06	620644.047655	63375.0	877617.0	1171795.0	1238705.0	13454352.0
Clump_Thickness	683.0	4.442167e+00	2.820761	1.0	2.0	4.0	6.0	10.0
Cell_Size	683.0	3.150805e+00	3.065145	1.0	1.0	1.0	5.0	10.0
Cell_Shape	683.0	3.215227e+00	2.988581	1.0	1.0	1.0	5.0	10.0
Marginal_Adhesion	683.0	2.830161e+00	2.864562	1.0	1.0	1.0	4.0	10.0
Single_Epithelial_Cell_Size	683.0	3.234261e+00	2.223085	1.0	2.0	2.0	4.0	10.0
Bare_Nuclei	683.0	3.544656e+00	3.643857	1.0	1.0	1.0	6.0	10.0
Bland_Chromatin	683.0	3.445095e+00	2.449697	1.0	2.0	3.0	5.0	10.0
Normal_Nucleoli	683.0	2.869693e+00	3.052666	1.0	1.0	1.0	4.0	10.0
Mitoses	683.0	1.603221e+00	1.732674	1.0	1.0	1.0	1.0	10.0
Class	683.0	3.499268e-01	0.477296	0.0	0.0	0.0	1.0	1.0

In [11]:

data.columns.tolist()

Out[11]:

['code',
 'Clump_Thickness',
 'Cell_Size',
 'Cell_Shape',
 'Marginal_Adhesion',
 'Single_Epithelial_Cell_Size',
 'Bare_Nuclei',
 'Bland_Chromatin',
 'Normal_Nucleoli',
 'Mitoses',
 'Class']

In [32]:

l=data.columns.tolist()[1:10]
l

Out[32]:

['Clump_Thickness',
 'Cell_Size',
 'Cell_Shape',
 'Marginal_Adhesion',
 'Single_Epithelial_Cell_Size',
 'Bare_Nuclei',
 'Bland_Chromatin',
 'Normal_Nucleoli',
 'Mitoses']

In [9]:

data.shape

Out[9]:

(683, 11)

유방암 환자 비율 확인¶

In [8]:

data['Class'].value_counts()

Out[8]:

0    444
1    239
Name: Class, dtype: int64

특성치(설명변수)와 레이블(종속변수) 나누기¶

특성치 (설명변수) 분류하기¶

방법1 : 변수명으로 분리하기 (=column 명)¶

In [36]:

X1=data[[ 'Clump_Thickness', 'Cell_Size', 'Cell_Shape', 'Marginal_Adhesion', 'Single_Epithelial_Cell_Size', 'Bare_Nuclei',
 'Bland_Chromatin', 'Normal_Nucleoli', 'Mitoses']]
X1

Out[36]:

	Clump_Thickness	Cell_Size	Cell_Shape	Marginal_Adhesion	Single_Epithelial_Cell_Size	Bare_Nuclei	Bland_Chromatin	Normal_Nucleoli	Mitoses
0	5	1	1	1	2	1	3	1	1
1	5	4	4	5	7	10	3	2	1
2	3	1	1	1	2	2	3	1	1
3	6	8	8	1	3	4	3	7	1
4	4	1	1	3	2	1	3	1	1
...	...	...	...	...	...	...	...	...	...
678	3	1	1	1	3	2	1	1	1
679	2	1	1	1	2	1	1	1	1
680	5	10	10	3	7	3	8	10	2
681	4	8	6	4	3	4	10	6	1
682	4	8	8	5	4	5	10	4	1

683 rows × 9 columns

In [37]:

X1_1=data[data.columns.tolist()[1:10]]
X1_1

Out[37]:

	Clump_Thickness	Cell_Size	Cell_Shape	Marginal_Adhesion	Single_Epithelial_Cell_Size	Bare_Nuclei	Bland_Chromatin	Normal_Nucleoli	Mitoses
0	5	1	1	1	2	1	3	1	1
1	5	4	4	5	7	10	3	2	1
2	3	1	1	1	2	2	3	1	1
3	6	8	8	1	3	4	3	7	1
4	4	1	1	3	2	1	3	1	1
...	...	...	...	...	...	...	...	...	...
678	3	1	1	1	3	2	1	1	1
679	2	1	1	1	2	1	1	1	1
680	5	10	10	3	7	3	8	10	2
681	4	8	6	4	3	4	10	6	1
682	4	8	8	5	4	5	10	4	1

683 rows × 9 columns

방법2 : Column indexing¶

In [38]:

X2=data[data.columns[1:10]]
X2

Out[38]:

	Clump_Thickness	Cell_Size	Cell_Shape	Marginal_Adhesion	Single_Epithelial_Cell_Size	Bare_Nuclei	Bland_Chromatin	Normal_Nucleoli	Mitoses
0	5	1	1	1	2	1	3	1	1
1	5	4	4	5	7	10	3	2	1
2	3	1	1	1	2	2	3	1	1
3	6	8	8	1	3	4	3	7	1
4	4	1	1	3	2	1	3	1	1
...	...	...	...	...	...	...	...	...	...
678	3	1	1	1	3	2	1	1	1
679	2	1	1	1	2	1	1	1	1
680	5	10	10	3	7	3	8	10	2
681	4	8	6	4	3	4	10	6	1
682	4	8	8	5	4	5	10	4	1

683 rows × 9 columns

방법3: loc method 이용¶

• 유의사항 : 인덱싱 차이 유의 - 마지막 범위까지 포함
• EX) 1:10 => 1열부터 10열까지

In [40]:

X3=data.loc[:,"Clump_Thickness":"Mitoses"]
X3

Out[40]:

	Clump_Thickness	Cell_Size	Cell_Shape	Marginal_Adhesion	Single_Epithelial_Cell_Size	Bare_Nuclei	Bland_Chromatin	Normal_Nucleoli	Mitoses
0	5	1	1	1	2	1	3	1	1
1	5	4	4	5	7	10	3	2	1
2	3	1	1	1	2	2	3	1	1
3	6	8	8	1	3	4	3	7	1
4	4	1	1	3	2	1	3	1	1
...	...	...	...	...	...	...	...	...	...
678	3	1	1	1	3	2	1	1	1
679	2	1	1	1	2	1	1	1	1
680	5	10	10	3	7	3	8	10	2
681	4	8	6	4	3	4	10	6	1
682	4	8	8	5	4	5	10	4	1

683 rows × 9 columns

iloc 사용해도 결과는 동일¶

In [41]:

data.iloc[:,1:10]

Out[41]:

	Clump_Thickness	Cell_Size	Cell_Shape	Marginal_Adhesion	Single_Epithelial_Cell_Size	Bare_Nuclei	Bland_Chromatin	Normal_Nucleoli	Mitoses
0	5	1	1	1	2	1	3	1	1
1	5	4	4	5	7	10	3	2	1
2	3	1	1	1	2	2	3	1	1
3	6	8	8	1	3	4	3	7	1
4	4	1	1	3	2	1	3	1	1
...	...	...	...	...	...	...	...	...	...
678	3	1	1	1	3	2	1	1	1
679	2	1	1	1	2	1	1	1	1
680	5	10	10	3	7	3	8	10	2
681	4	8	6	4	3	4	10	6	1
682	4	8	8	5	4	5	10	4	1

683 rows × 9 columns

레이블(종속변수) 분류하기¶

In [42]:

y=data[['Class']]
y

Out[42]:

	Class
0	0
1	0
2	0
3	0
4	0
...	...
678	0
679	0
680	1
681	1
682	1

683 rows × 1 columns

train/ test set 나누기¶

-주요 파라미터

1. stratify (층화) : 범주 비율에 맞게 추출
2. random_state=42 : 난수 규칙, (같은 결과 확인)
3. test_size=0.25 (기본값) : test set 비율
4. shuffle (기본값 : True) : shuffle 여부

sklearn import¶

In [44]:

from sklearn.model_selection import train_test_split

data 분할하기¶

In [45]:

X_train,X_test,y_train,y_test=train_test_split(X1,y,stratify=y,random_state=42)

train/ test set 비율¶

In [52]:

len(X_test)/(len(X_test)+len(X_train))

Out[52]:

0.25036603221083453

In [46]:

y_train.mean()

Out[46]:

Class    0.349609
dtype: float64

In [55]:

y_test.var()

Out[55]:

Class    0.229102
dtype: float64

In [47]:

y_test.mean()

Out[47]:

Class    0.350877
dtype: float64

In [54]:

y_train.var()

Out[54]:

Class    0.227828
dtype: float64

data set 정규화 (Min-Man, 표준화)¶

※ train set에 fitting 한 후 test set도 정규화 할 것

• 순서 : .fit -> .transform

라이브러리 Import¶

In [59]:

from sklearn.preprocessing import MinMaxScaler
from sklearn.preprocessing import StandardScaler
# 정규화 하기 위한 Class 선언

scaler_minmax=MinMaxScaler()
scaler_standard=StandardScaler()

정규화 1: Min-Max 정규화 : numpy array로 반환¶

In [68]:

# X_train set에 fit (train 셋의 최대 최소를 기준으로 0~1사이 값으로 변환)
scaler_minmax.fit(X_train)
X_scaled_minmax_train=scaler_minmax.transform(X_train)
# test set은 별도의 fit 과정 필요 없음 (X_train set에 맞춰 fitting 되어있음)
X_scaled_minmax_test=scaler_minmax.transform(X_test)

정규화 2 : 표준화 (정규분포에 fitting) :numpy array로 반환¶

In [79]:

# X_train set에 fit (train 셋을 정규 분포에 근사)
scaler_standard.fit(X_train)
X_scaled_standard_train=scaler_standard.transform(X_train)
# test set은 별도의 fit 과정 필요 없음 (X_train set에 맞춰 fitting 되어있음)
X_scaled_standard_test=scaler_standard.transform(X_test)

정규화 결과 확인하기¶

In [75]:

# min-max 정규화 - train set
pd.DataFrame(X_scaled_minmax_train).describe()

Out[75]:

	0	1	2	3	4	5	6	7	8
count	512.000000	512.000000	512.000000	512.000000	512.000000	512.000000	512.000000	512.000000	512.000000
mean	0.372830	0.231988	0.242839	0.205078	0.241319	0.285590	0.269314	0.199002	0.067491
std	0.317836	0.334781	0.332112	0.319561	0.242541	0.404890	0.265289	0.331503	0.190373
min	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000
25%	0.111111	0.000000	0.000000	0.000000	0.111111	0.000000	0.111111	0.000000	0.000000
50%	0.333333	0.000000	0.000000	0.000000	0.111111	0.000000	0.222222	0.000000	0.000000
75%	0.555556	0.361111	0.444444	0.333333	0.333333	0.583333	0.444444	0.222222	0.000000
max	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000

In [76]:

# 표준화 - train set
pd.DataFrame(X_scaled_standard_train).describe()

Out[76]:

	0	1	2	3	4	5	6	7	8
count	5.120000e+02	5.120000e+02	5.120000e+02	5.120000e+02	5.120000e+02	5.120000e+02	5.120000e+02	5.120000e+02	5.120000e+02
mean	-1.548241e-16	-1.543904e-16	-1.353084e-16	1.149254e-16	5.767956e-17	1.674008e-16	-2.775558e-17	-3.642919e-17	6.938894e-18
std	1.000978e+00	1.000978e+00	1.000978e+00	1.000978e+00	1.000978e+00	1.000978e+00	1.000978e+00	1.000978e+00	1.000978e+00
min	-1.174173e+00	-6.936309e-01	-7.319088e-01	-6.423777e-01	-9.959361e-01	-7.060427e-01	-1.016165e+00	-6.008881e-01	-3.548677e-01
25%	-8.242452e-01	-6.936309e-01	-7.319088e-01	-6.423777e-01	-5.373756e-01	-7.060427e-01	-5.969255e-01	-6.008881e-01	-3.548677e-01
50%	-1.243886e-01	-6.936309e-01	-7.319088e-01	-6.423777e-01	-5.373756e-01	-7.060427e-01	-1.776856e-01	-6.008881e-01	-3.548677e-01
75%	5.754680e-01	3.860715e-01	6.076347e-01	4.017410e-01	3.797454e-01	7.360871e-01	6.607941e-01	7.011454e-02	-3.548677e-01
max	1.975181e+00	2.296314e+00	2.282064e+00	2.489978e+00	3.131108e+00	1.766180e+00	2.756993e+00	2.418624e+00	4.903108e+00

In [78]:

pd.DataFrame(X_scaled_minmax_test).describe()

Out[78]:

	0	1	2	3	4	5	6	7	8
count	171.000000	171.000000	171.000000	171.000000	171.000000	171.000000	171.000000	171.000000	171.000000
mean	0.411306	0.259909	0.256010	0.198181	0.269006	0.274204	0.278752	0.233918	0.065627
std	0.298847	0.357544	0.332700	0.315307	0.259557	0.405891	0.292578	0.360958	0.199372
min	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000	0.000000
25%	0.222222	0.000000	0.000000	0.000000	0.111111	0.000000	0.000000	0.000000	0.000000
50%	0.444444	0.000000	0.111111	0.000000	0.111111	0.000000	0.222222	0.000000	0.000000
75%	0.555556	0.444444	0.444444	0.222222	0.388889	0.444444	0.444444	0.388889	0.000000
max	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000	1.000000

In [77]:

pd.DataFrame(X_scaled_standard_test).describe()

Out[77]:

	0	1	2	3	4	5	6	7	8
count	171.000000	171.000000	171.000000	171.000000	171.000000	171.000000	171.000000	171.000000	171.000000
mean	0.121175	0.083483	0.039700	-0.021605	0.114263	-0.028149	0.035612	0.105430	-0.009802
std	0.941174	1.069038	1.002747	0.987654	1.071204	1.003453	1.103943	1.089918	1.048292
min	-1.174173	-0.693631	-0.731909	-0.642378	-0.995936	-0.706043	-1.016165	-0.600888	-0.354868
25%	-0.474317	-0.693631	-0.731909	-0.642378	-0.537376	-0.706043	-1.016165	-0.600888	-0.354868
50%	0.225540	-0.693631	-0.397023	-0.642378	-0.537376	-0.706043	-0.177686	-0.600888	-0.354868
75%	0.575468	0.635234	0.607635	0.053701	0.609026	0.392723	0.660794	0.573367	-0.354868
max	1.975181	2.296314	2.282064	2.489978	3.131108	1.766180	2.756993	2.418624	4.903108

모델 학습¶

sklearn 라이브러리 import¶

In [81]:

from sklearn.linear_model import LogisticRegression
# Class 선언
model=LogisticRegression()

모델 학습 (train set 이용하기)¶

In [82]:

model.fit(X_scaled_minmax_train,y_train) #모델에 학습

C:\ProgramData\Anaconda3\lib\site-packages\sklearn\utils\validation.py:72: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().
  return f(**kwargs)

Out[82]:

LogisticRegression()

train set의 예측치/정확도 구하기¶

• 예측치 : model.predict
• 정확도(Accuarcy) : model.score(x_train, y_train)

In [89]:

pred_train=model.predict(X_scaled_minmax_train)
print( "Logistic 회귀 모델의 정확도 : " ,model.score(X_scaled_minmax_train,y_train))

Logistic 회귀 모델의 정확도 :  0.97265625

모델의 예측 확률¶

In [91]:

model.predict_proba(X_scaled_minmax_train)

Out[91]:

array([[0.98101387, 0.01898613],
       [0.7681914 , 0.2318086 ],
       [0.96643115, 0.03356885],
       ...,
       [0.11344041, 0.88655959],
       [0.98740488, 0.01259512],
       [0.99046984, 0.00953016]])

회귀 모델의 상수 : coef_¶

In [96]:

model.coef_.tolist()[0]

Out[96]:

[2.5375183632972074,
 1.7450083798212197,
 1.9517785814607636,
 1.381752935362637,
 1.1194519302008699,
 3.2896703792679056,
 1.403280974947181,
 1.2145170680225785,
 1.1424321977229857]

회귀 모델의 y절편 : intercept_¶

In [87]:

model.intercept_

Out[87]:

array([-4.92402106])

변수별 상수 구하기¶

In [150]:

dict_coef={var:coef for var,coef in zip(data.columns.tolist()[1:10],model.coef_.tolist()[0])}
dict_coef
for k,v in dict_coef.items():
    print("변수명:",k, "  / 계수:",v)

변수명: Clump_Thickness   / 계수: 2.5375183632972074
변수명: Cell_Size   / 계수: 1.7450083798212197
변수명: Cell_Shape   / 계수: 1.9517785814607636
변수명: Marginal_Adhesion   / 계수: 1.381752935362637
변수명: Single_Epithelial_Cell_Size   / 계수: 1.1194519302008699
변수명: Bare_Nuclei   / 계수: 3.2896703792679056
변수명: Bland_Chromatin   / 계수: 1.403280974947181
변수명: Normal_Nucleoli   / 계수: 1.2145170680225785
변수명: Mitoses   / 계수: 1.1424321977229857

test set의 예측치/ 정확도 확인하기¶

• fitting 과정 필요 없음

In&nbnbsp;[92]:

pred_test=model.predict(X_scaled_minmax_test)
model.score(X_scaled_minmax_test,y_test)

Out[92]:

0.9590643274853801

결과 확인하기 : confusion matrix¶

이미지 출처 : 나무 위키¶

1. 정확도 (Accuracy) : 예측이 정확할 확률 (참긍정+참부정)/(전체 예측)

2. 정밀도 (Precision) : 예측이 긍정일 때 실제로 긍정일 확률 (참긍정)/(참긍정+거짓긍정)

3. 민감도(=재현율)/ Sensitivity(=Recall) : 실제로 긍정일 때 예측도 긍정일 확률 (참금정)/ (참긍정+거짓부정)

4. 특이도(Specificity) : 실제로 부정일 때 예측도 부정일 확률 (참 부정)/(참부정 + 거짓긍정)

5. F1 : 정밀도와 민감도의 조화평균
   -> 정밀도와 재현율 중 하나가 높아지면 다른 하나가 낮아지는 상황을 고려하여 보정하기 위해 활용

혼동행렬 출력하기¶

train set¶

In [107]:

from sklearn.metrics import confusion_matrix
confusion_train=confusion_matrix(y_train,pred_train)
print(confusion_train)

[[328   5]
 [  9 170]]

test set¶

In [109]:

confusion_train=confusion_matrix(y_test,pred_test)
print(confusion_train)

[[106   5]
 [  2  58]]

평가지표 확인하기 : report 출력하기¶

train set¶

In [112]:

from sklearn.metrics import classification_report
cfreport_train=classification_report(y_train,pred_train)
print(cfreport_train)

              precision    recall  f1-score   support

           0       0.97      0.98      0.98       333
           1       0.97      0.95      0.96       179

    accuracy                           0.97       512
   macro avg       0.97      0.97      0.97       512
weighted avg       0.97      0.97      0.97       512

test set¶

In [114]:

cfreport_test=classification_report(y_test,pred_test)
print(cfreport_test)

              precision    recall  f1-score   support

           0       0.98      0.95      0.97       111
           1       0.92      0.97      0.94        60

    accuracy                           0.96       171
   macro avg       0.95      0.96      0.96       171
weighted avg       0.96      0.96      0.96       171

ROC Curve 확인하기¶

• 참고자료 출처 : https://angeloyeo.github.io/2020/08/05/ROC.html

library import¶

In [115]:

from sklearn.metrics import roc_curve, auc

In [116]:

from sklearn import metrics

AUC 구하기¶

In [117]:

false_positive_rate, true_positive_rate, thresholds=roc_curve(y_test,model.decision_function(X_scaled_minmax_test))
roc_auc=metrics.roc_auc_score(y_test,model.decision_function(X_scaled_minmax_test))

Out[117]:

0.9923423423423423

In [118]:

roc_auc

Out[118]:

0.9923423423423423

• 결정함수 : 부호를 통해 값 확인 가능

In [120]:

model.decision_function(X_scaled_minmax_test)[1:4]

Out[120]:

array([-3.5070115 , -2.81670291, -3.8601789 ])

ROC Curve 그리기¶

In [126]:

import matplotlib.pyplot as plt
# title 입력
plt.title("ROC Curve")
# x축 
plt.xlabel("False positive rate(1-Specficity)")
# y축
plt.ylabel("True positive rate(Sensitivity)")
# roc curve 그리기
# 색상 지정 : "b"
plt.plot(false_positive_rate,true_positive_rate,"b",label="Model(AUC= %0.2f)"%roc_auc)
# 기준선 그리기
# 선타입/마커 지정
plt.plot([0,1],[1,1],"y--") # (0,1) /(1,1) 지나는 직선, 색상은 yellow
plt.plot([0,1],[0,1],"r--")
plt.legend(loc="lower right")
plt.show()

참고자료 : matplotlib 선/마커 지정하기¶

• 출처 : https://wikidocs.net/92083#_2

선/마커 표시 형식 예시

선/마커 동시 지정하기

설정 방법

예측 결과 정리¶

train set 예측확률 Dataframe에 저장하기¶

In [137]:

prob_train=model.predict_proba(X_scaled_minmax_train) # 0,1에 대한 각각의 확률 예측값


y_train[['y_pred']]=pred_train #trainset에 의한 예측 결과 저장

y_train[['y_prob0','y_prob1']]=prob_train
y_train

Out[137]:

	Class	y_pred	y_prob0	y_prob1
131	0	0	0.981014	0.018986
6	0	0	0.768191	0.231809
0	0	0	0.966431	0.033569
269	0	0	0.988880	0.011120
56	1	1	0.203161	0.796839
...	...	...	...	...
515	1	1	0.021270	0.978730
216	1	0	0.895961	0.104039
312	1	1	0.113440	0.886560
11	0	0	0.987405	0.012595
268	0	0	0.990470	0.009530

512 rows × 4 columns

In [132]:

pd.DataFrame(prob_train)

Out[132]:

	0	1
0	0.981014	0.018986
1	0.768191	0.231809
2	0.966431	0.033569
3	0.988880	0.011120
4	0.203161	0.796839
...	...	...
507	0.021270	0.978730
508	0.895961	0.104039
509	0.113440	0.886560
510	0.987405	0.012595
511	0.990470	0.009530

512 rows × 2 columns

test set 예측확률 Dataframe에 저장하기¶

In [138]:

prob_test=model.predict_proba(X_scaled_minmax_test)
y_test[['y_pred']]=pred_test
y_test[['y_prob0','y_prob1']]=prob_test
y_test

Out[138]:

	Class	y_pred	y_prob0	y_prob1
541	0	0	0.955893	0.044107
549	0	0	0.970887	0.029113
318	0	0	0.943572	0.056428
183	0	0	0.979370	0.020630
478	1	1	0.001305	0.998695
...	...	...	...	...
425	1	1	0.006201	0.993799
314	1	1	0.067440	0.932560
15	1	1	0.436887	0.563113
510	0	0	0.983410	0.016590
351	0	0	0.987405	0.012595

171 rows × 4 columns

전체 결과 Dataframe으로 병합 후 csv 파일로 저장하기¶

In [139]:

X_test

Out[139]:

	Clump_Thickness	Cell_Size	Cell_Shape	Marginal_Adhesion	Single_Epithelial_Cell_Size	Bare_Nuclei	Bland_Chromatin	Normal_Nucleoli	Mitoses
541	5	2	2	2	1	1	2	1	1
549	4	1	1	1	2	1	3	2	1
318	5	2	2	2	2	1	2	2	1
183	1	2	3	1	2	1	3	1	1
478	5	10	10	10	6	10	6	5	2
...	...	...	...	...	...	...	...	...	...
425	10	4	3	10	4	10	10	1	1
314	8	10	3	2	6	4	3	10	1
15	7	4	6	4	6	1	4	3	1
510	3	1	1	2	2	1	1	1	1
351	2	1	1	1	2	1	2	1	1

171 rows × 9 columns

In [140]:

y_test

Out[140]:

	Class	y_pred	y_prob0	y_prob1
541	0	0	0.955893	0.044107
549	0	0	0.970887	0.029113
318	0	0	0.943572	0.056428
183	0	0	0.979370	0.020630
478	1	1	0.001305	0.998695
...	...	...	...	...
425	1	1	0.006201	0.993799
314	1	1	0.067440	0.932560
15	1	1	0.436887	0.563113
510	0	0	0.983410	0.016590
351	0	0	0.987405	0.012595

171 rows × 4 columns

In [147]:

# pd.concat : dataframe 합치기
# axis=0 (기본값) : 열을 추가 (세로 병합)
# axis=1 : 변수를 합침 (가로 병합)
# Total_test=pd.concat([X_test,y_test],axis-1) 
Total_test=pd.concat([X_test,y_test], axis=1) 
# csv로 내보내기
Total_test.to_csv("classification_logistic.csv")

In [149]:

Total_test.head(20)

Out[149]:

	Clump_Thickness	Cell_Size	Cell_Shape	Marginal_Adhesion	Single_Epithelial_Cell_Size	Bare_Nuclei	Bland_Chromatin	Normal_Nucleoli	Mitoses	Class	y_pred	y_prob0	y_prob1
541	5	2	2	2	1	1	2	1	1	0	0	0.955893	0.044107
549	4	1	1	1	2	1	3	2	1	0	0	0.970887	0.029113
318	5	2	2	2	2	1	2	2	1	0	0	0.943572	0.056428
183	1	2	3	1	2	1	3	1	1	0	0	0.979370	0.020630
478	5	10	10	10	6	10	6	5	2	1	1	0.001305	0.998695
65	5	3	4	1	8	10	4	9	1	1	1	0.049745	0.950255
430	2	1	1	1	2	1	1	1	1	0	0	0.989204	0.010796
17	4	1	1	1	2	1	3	1	1	0	0	0.974468	0.025532
443	5	1	2	1	2	1	1	1	1	0	0	0.969380	0.030620
77	2	1	1	1	3	1	2	1	1	0	0	0.985760	0.014240
212	6	10	7	7	6	4	8	10	2	1	1	0.009908	0.990092
95	5	1	1	1	2	1	3	1	1	0	0	0.966431	0.033569
659	4	1	4	1	2	1	1	1	1	0	0	0.964539	0.035461
143	1	1	1	1	3	2	2	1	1	0	0	0.984538	0.015462
383	4	1	1	1	2	1	1	1	1	0	0	0.981179	0.018821
560	5	1	2	1	2	1	3	1	1	0	0	0.958638	0.041362
634	3	1	1	2	3	4	1	1	1	0	0	0.945899	0.054101
311	3	2	2	1	2	1	2	3	1	0	0	0.967679	0.032321
451	10	6	6	2	4	10	9	7	1	1	1	0.003909	0.996091
117	3	2	1	1	2	2	3	1	1	0	0	0.966576	0.033424

pd.concat axis=1,axis=0 차이¶

In [146]:

pd.concat([X_test,y_test], axis=0) 

Out[146]:

	Clump_Thickness	Cell_Size	Cell_Shape	Marginal_Adhesion	Single_Epithelial_Cell_Size	Bare_Nuclei	Bland_Chromatin	Normal_Nucleoli	Mitoses	Class	y_pred	y_prob0	y_prob1
541	5.0	2.0	2.0	2.0	1.0	1.0	2.0	1.0	1.0	NaN	NaN	NaN	NaN
549	4.0	1.0	1.0	1.0	2.0	1.0	3.0	2.0	1.0	NaN	NaN	NaN	NaN
318	5.0	2.0	2.0	2.0	2.0	1.0	2.0	2.0	1.0	NaN	NaN	NaN	NaN
183	1.0	2.0	3.0	1.0	2.0	1.0	3.0	1.0	1.0	NaN	NaN	NaN	NaN
478	5.0	10.0	10.0	10.0	6.0	10.0	6.0	5.0	2.0	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...	...	...	...	...	...	...	...
425	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.0	1.0	0.006201	0.993799
314	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.0	1.0	0.067440	0.932560
15	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.0	1.0	0.436887	0.563113
510	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.0	0.0	0.983410	0.016590
351	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	0.0	0.0	0.987405	0.012595

342 rows × 13 columns

In [ ]: