코딩/머신러닝과 딥러닝
LightGBM
Song1234
2024. 7. 17. 11:14
1. credit 데이터셋¶
In [169]:
import numpy as np
import pandas as pd
import seaborn as sns
import matplotlib.pyplot as plt
In [170]:
credit_df = pd.read_csv('/content/drive/MyDrive/KDT/6. 머신러닝과 딥러닝/Data/credit.csv')
credit_df
Out[170]:
| ID | Customer_ID | Name | Age | SSN | Occupation | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | ... | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Payment_of_Min_Amount | Total_EMI_per_month | Amount_invested_monthly | Payment_Behaviour | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 0x1602 | CUS_0xd40 | Aaron Maashoh | 23 | 821-00-0265 | Scientist | 19114.12 | 3 | 4 | 3 | ... | 4.0 | 809.98 | 26.822620 | 22 Years and 1 Months | No | 49.574949 | 80.41529544 | High_spent_Small_value_payments | 312.494089 | Good |
| 1 | 0x160e | CUS_0x21b1 | Rick Rothackerj | 28_ | 004-07-5839 | _______ | 34847.84 | 2 | 4 | 6 | ... | 2.0 | 605.03 | 24.464031 | 26 Years and 7 Months | No | 18.816215 | 104.2918252 | Low_spent_Small_value_payments | 470.690627 | Standard |
| 2 | 0x161a | CUS_0x2dbc | Langep | 34 | 486-85-3974 | _______ | 143162.64 | 1 | 5 | 8 | ... | 3.0 | 1303.01 | 28.616735 | 17 Years and 9 Months | No | 246.992320 | 168.4137027 | !@9#%8 | 1043.315978 | Good |
| 3 | 0x1626 | CUS_0xb891 | Jasond | 54 | 072-31-6145 | Entrepreneur | 30689.89 | 2 | 5 | 4 | ... | 4.0 | 632.46 | 26.544229 | 17 Years and 3 Months | No | 16.415452 | 81.22885871 | Low_spent_Large_value_payments | 433.604773 | Standard |
| 4 | 0x1632 | CUS_0x1cdb | Deepaa | 21 | 615-06-7821 | Developer | 35547.71_ | 7 | 5 | 5 | ... | 4.0 | 943.86 | 39.797764 | 30 Years and 8 Months | Yes | 0.000000 | 276.7253943 | !@9#%8 | 288.605522 | Standard |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 12495 | 0x25fb6 | CUS_0x372c | Lucia Mutikanik | 18 | 340-85-7301 | Lawyer | 42903.79 | 0 | 4 | 6 | ... | 1.0 | 1079.48 | 27.289440 | 28 Years and 1 Months | No | 50894.000000 | 78.51494451 | High_spent_Small_value_payments | 493.341182 | Good |
| 12496 | 0x25fc2 | CUS_0xf16 | Maria Sheahanb | 44 | #F%$D@*&8 | Media_Manager | 16680.35 | 1 | 1 | 5 | ... | 4.0 | 897.16 | 39.868572 | NaN | NM | 41.113561 | 52.95197782 | High_spent_Small_value_payments | 318.737378 | Good |
| 12497 | 0x25fce | CUS_0xaf61 | Chris Wickhamm | 49 | 133-16-7738 | Writer | 37188.1 | 1 | 4 | 5 | ... | 3.0 | 620.64 | 39.080823 | 29 Years and 9 Months | No | 84.205949 | 223.8750182 | Low_spent_Small_value_payments | 291.619866 | Good |
| 12498 | 0x25fda | CUS_0x8600 | Sarah McBridec | 28 | 031-35-0942 | Architect | 20002.88 | 10 | 8 | 29 | ... | 9.0 | 3571.7_ | 22.895966 | 5 Years and 8 Months | Yes | 60.964772 | 43.37067007 | High_spent_Large_value_payments | 328.655224 | Poor |
| 12499 | 0x25fe6 | CUS_0x942c | Nicks | 24 | 078-73-5990 | Mechanic | 39628.99 | 4 | 6 | 7 | ... | 3.0 | 502.38 | 32.991333 | 31 Years and 3 Months | No | 35.104023 | 401.1964806 | Low_spent_Small_value_payments | 189.641080 | Poor |
12500 rows × 24 columns
In [171]:
credit_df.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 12500 entries, 0 to 12499
Data columns (total 24 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 ID 12500 non-null object
1 Customer_ID 12500 non-null object
2 Name 11273 non-null object
3 Age 12500 non-null object
4 SSN 12500 non-null object
5 Occupation 12500 non-null object
6 Annual_Income 12500 non-null object
7 Num_Bank_Accounts 12500 non-null int64
8 Num_Credit_Card 12500 non-null int64
9 Interest_Rate 12500 non-null int64
10 Num_of_Loan 12500 non-null object
11 Type_of_Loan 11074 non-null object
12 Delay_from_due_date 12500 non-null int64
13 Num_of_Delayed_Payment 11657 non-null object
14 Num_Credit_Inquiries 12264 non-null float64
15 Outstanding_Debt 12500 non-null object
16 Credit_Utilization_Ratio 12500 non-null float64
17 Credit_History_Age 11387 non-null object
18 Payment_of_Min_Amount 12500 non-null object
19 Total_EMI_per_month 12500 non-null float64
20 Amount_invested_monthly 11935 non-null object
21 Payment_Behaviour 12500 non-null object
22 Monthly_Balance 12366 non-null float64
23 Credit_Score 12500 non-null object
dtypes: float64(4), int64(4), object(16)
memory usage: 2.3+ MB
- ID: 고유 식별자
- Customer_ID: 고객 ID
- Name: 이름
- Age: 나이
- SSN: 주민등록번호
- Occupation: 직업
- Annual_Income: 연간 소득
- Num_Bank_Accounts: 은행 계좌 수
- Num_Credit_Card: 신용 카드 수
- Interest_Rate: 이자율
- Num_of_Loan: 대출 수
- Type_of_Loan: 대출 유형
- Delay_from_due_date: 마감일로부터 연체 기간
- Num_of_Delayed_Payment: 연체된 결제 수
- Num_Credit_Inquiries: 신용조회 수
- Outstanding_Debt: 미상환 잔금
- Credit_Utilization_Ratio: 신용카드 사용률
- Credit_History_Age: 카드 사용 기간
- Payment_of_Min_Amount: 리볼빙 여부
- Total_EMI_per_month: 월별 총 지출 금액
- Amount_invested_monthly: 매월 투자 금액
- Payment_Behaviour: 지불 행동
- Monthly_Balance: 월별 잔고
- Credit_Score: 신용 점수
In [172]:
credit_df.drop(['ID', 'Customer_ID', 'Name', 'SSN'], axis=1, inplace=True)
In [173]:
credit_df.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 12500 entries, 0 to 12499
Data columns (total 20 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Age 12500 non-null object
1 Occupation 12500 non-null object
2 Annual_Income 12500 non-null object
3 Num_Bank_Accounts 12500 non-null int64
4 Num_Credit_Card 12500 non-null int64
5 Interest_Rate 12500 non-null int64
6 Num_of_Loan 12500 non-null object
7 Type_of_Loan 11074 non-null object
8 Delay_from_due_date 12500 non-null int64
9 Num_of_Delayed_Payment 11657 non-null object
10 Num_Credit_Inquiries 12264 non-null float64
11 Outstanding_Debt 12500 non-null object
12 Credit_Utilization_Ratio 12500 non-null float64
13 Credit_History_Age 11387 non-null object
14 Payment_of_Min_Amount 12500 non-null object
15 Total_EMI_per_month 12500 non-null float64
16 Amount_invested_monthly 11935 non-null object
17 Payment_Behaviour 12500 non-null object
18 Monthly_Balance 12366 non-null float64
19 Credit_Score 12500 non-null object
dtypes: float64(4), int64(4), object(12)
memory usage: 1.9+ MB
In [174]:
credit_df['Credit_Score'].value_counts()
Out[174]:
Credit_Score
Standard 6943
Poor 3582
Good 1975
Name: count, dtype: int64In [175]:
credit_df['Credit_Score'] = credit_df['Credit_Score'].replace({'Poor':0, 'Standard':1, 'Good':2})
credit_df.head()
Out[175]:
| Age | Occupation | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Type_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Payment_of_Min_Amount | Total_EMI_per_month | Amount_invested_monthly | Payment_Behaviour | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 23 | Scientist | 19114.12 | 3 | 4 | 3 | 4 | Auto Loan, Credit-Builder Loan, Personal Loan,... | 3 | 7 | 4.0 | 809.98 | 26.822620 | 22 Years and 1 Months | No | 49.574949 | 80.41529544 | High_spent_Small_value_payments | 312.494089 | 2 |
| 1 | 28_ | _______ | 34847.84 | 2 | 4 | 6 | 1 | Credit-Builder Loan | 3 | 4 | 2.0 | 605.03 | 24.464031 | 26 Years and 7 Months | No | 18.816215 | 104.2918252 | Low_spent_Small_value_payments | 470.690627 | 1 |
| 2 | 34 | _______ | 143162.64 | 1 | 5 | 8 | 3 | Auto Loan, Auto Loan, and Not Specified | 5 | 8 | 3.0 | 1303.01 | 28.616735 | 17 Years and 9 Months | No | 246.992320 | 168.4137027 | !@9#%8 | 1043.315978 | 2 |
| 3 | 54 | Entrepreneur | 30689.89 | 2 | 5 | 4 | 1 | Not Specified | 0 | 6 | 4.0 | 632.46 | 26.544229 | 17 Years and 3 Months | No | 16.415452 | 81.22885871 | Low_spent_Large_value_payments | 433.604773 | 1 |
| 4 | 21 | Developer | 35547.71_ | 7 | 5 | 5 | 0 | NaN | 5 | NaN | 4.0 | 943.86 | 39.797764 | 30 Years and 8 Months | Yes | 0.000000 | 276.7253943 | !@9#%8 | 288.605522 | 1 |
In [176]:
credit_df.describe()
Out[176]:
| Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Delay_from_due_date | Num_Credit_Inquiries | Credit_Utilization_Ratio | Total_EMI_per_month | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|
| count | 12500.000000 | 12500.000000 | 12500.00000 | 12500.000000 | 12264.000000 | 12500.000000 | 12500.000000 | 12366.000000 | 12500.000000 |
| mean | 17.275120 | 21.647680 | 69.46520 | 21.051440 | 24.591650 | 32.291949 | 1303.781040 | 405.815391 | 0.871440 |
| std | 118.518214 | 123.789969 | 455.95698 | 14.859994 | 183.422458 | 5.084327 | 8118.261086 | 218.136964 | 0.654268 |
| min | 0.000000 | 1.000000 | 1.00000 | -5.000000 | 0.000000 | 20.992914 | 0.000000 | 0.088628 | 0.000000 |
| 25% | 3.000000 | 4.000000 | 8.00000 | 10.000000 | 2.000000 | 28.110034 | 29.128806 | 271.785749 | 0.000000 |
| 50% | 6.000000 | 5.000000 | 13.00000 | 18.000000 | 4.000000 | 32.297912 | 66.372879 | 337.169588 | 1.000000 |
| 75% | 7.000000 | 7.000000 | 20.00000 | 28.000000 | 8.000000 | 36.458660 | 149.904496 | 475.222487 | 1.000000 |
| max | 1779.000000 | 1479.000000 | 5788.00000 | 67.000000 | 2592.000000 | 49.564519 | 82122.000000 | 1602.040519 | 2.000000 |
In [177]:
sns.barplot(x='Payment_of_Min_Amount', y='Credit_Score', data=credit_df)
Out[177]:
<Axes: xlabel='Payment_of_Min_Amount', ylabel='Credit_Score'>
In [178]:
plt.figure(figsize=(20, 5))
sns.barplot(x='Occupation', y='Credit_Score', data=credit_df)
Out[178]:
<Axes: xlabel='Occupation', ylabel='Credit_Score'>
In [179]:
plt.figure(figsize=(12, 12))
sns.heatmap(credit_df.corr(numeric_only=True), cmap='coolwarm', vmin=-1, vmax=1, annot=True)
Out[179]:
<Axes: >
In [180]:
credit_df.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 12500 entries, 0 to 12499
Data columns (total 20 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Age 12500 non-null object
1 Occupation 12500 non-null object
2 Annual_Income 12500 non-null object
3 Num_Bank_Accounts 12500 non-null int64
4 Num_Credit_Card 12500 non-null int64
5 Interest_Rate 12500 non-null int64
6 Num_of_Loan 12500 non-null object
7 Type_of_Loan 11074 non-null object
8 Delay_from_due_date 12500 non-null int64
9 Num_of_Delayed_Payment 11657 non-null object
10 Num_Credit_Inquiries 12264 non-null float64
11 Outstanding_Debt 12500 non-null object
12 Credit_Utilization_Ratio 12500 non-null float64
13 Credit_History_Age 11387 non-null object
14 Payment_of_Min_Amount 12500 non-null object
15 Total_EMI_per_month 12500 non-null float64
16 Amount_invested_monthly 11935 non-null object
17 Payment_Behaviour 12500 non-null object
18 Monthly_Balance 12366 non-null float64
19 Credit_Score 12500 non-null int64
dtypes: float64(4), int64(5), object(11)
memory usage: 1.9+ MB
In [181]:
credit_df['Payment_Behaviour'].dtype
Out[181]:
dtype('O')In [182]:
for i in credit_df.columns:
if credit_df[i].dtype == 'O':
print(i)
Age
Occupation
Annual_Income
Num_of_Loan
Type_of_Loan
Num_of_Delayed_Payment
Outstanding_Debt
Credit_History_Age
Payment_of_Min_Amount
Amount_invested_monthly
Payment_Behaviour
In [183]:
for i in ['Age', 'Annual_Income', 'Num_of_Loan', 'Num_of_Delayed_Payment', 'Outstanding_Debt', 'Amount_invested_monthly']:
credit_df[i] = pd.to_numeric(credit_df[i].str.replace('_', ''))
In [184]:
credit_df.info()
<class 'pandas.core.frame.DataFrame'>
RangeIndex: 12500 entries, 0 to 12499
Data columns (total 20 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Age 12500 non-null int64
1 Occupation 12500 non-null object
2 Annual_Income 12500 non-null float64
3 Num_Bank_Accounts 12500 non-null int64
4 Num_Credit_Card 12500 non-null int64
5 Interest_Rate 12500 non-null int64
6 Num_of_Loan 12500 non-null int64
7 Type_of_Loan 11074 non-null object
8 Delay_from_due_date 12500 non-null int64
9 Num_of_Delayed_Payment 11657 non-null float64
10 Num_Credit_Inquiries 12264 non-null float64
11 Outstanding_Debt 12500 non-null float64
12 Credit_Utilization_Ratio 12500 non-null float64
13 Credit_History_Age 11387 non-null object
14 Payment_of_Min_Amount 12500 non-null object
15 Total_EMI_per_month 12500 non-null float64
16 Amount_invested_monthly 11935 non-null float64
17 Payment_Behaviour 12500 non-null object
18 Monthly_Balance 12366 non-null float64
19 Credit_Score 12500 non-null int64
dtypes: float64(8), int64(7), object(5)
memory usage: 1.9+ MB
In [185]:
credit_df['Credit_History_Age']
Out[185]:
0 22 Years and 1 Months
1 26 Years and 7 Months
2 17 Years and 9 Months
3 17 Years and 3 Months
4 30 Years and 8 Months
...
12495 28 Years and 1 Months
12496 NaN
12497 29 Years and 9 Months
12498 5 Years and 8 Months
12499 31 Years and 3 Months
Name: Credit_History_Age, Length: 12500, dtype: objectIn [186]:
# Credit_History_Age의 데이터를 개월로 변경
# 22 Years and 1 Months -> 22 * 12 + 1 = 265
credit_df['Credit_History_Age'] = credit_df['Credit_History_Age'].str.replace(' Months', '')
# 22 Years and 1
credit_df['Credit_History_Age'] = pd.to_numeric(credit_df['Credit_History_Age'].str.split(' Years and ', expand=True)[0])*12 + pd.to_numeric(credit_df['Credit_History_Age'].str.split(' Years and ', expand=True)[1])
credit_df.head()
Out[186]:
| Age | Occupation | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Type_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Payment_of_Min_Amount | Total_EMI_per_month | Amount_invested_monthly | Payment_Behaviour | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 23 | Scientist | 19114.12 | 3 | 4 | 3 | 4 | Auto Loan, Credit-Builder Loan, Personal Loan,... | 3 | 7.0 | 4.0 | 809.98 | 26.822620 | 265.0 | No | 49.574949 | 80.415295 | High_spent_Small_value_payments | 312.494089 | 2 |
| 1 | 28 | _______ | 34847.84 | 2 | 4 | 6 | 1 | Credit-Builder Loan | 3 | 4.0 | 2.0 | 605.03 | 24.464031 | 319.0 | No | 18.816215 | 104.291825 | Low_spent_Small_value_payments | 470.690627 | 1 |
| 2 | 34 | _______ | 143162.64 | 1 | 5 | 8 | 3 | Auto Loan, Auto Loan, and Not Specified | 5 | 8.0 | 3.0 | 1303.01 | 28.616735 | 213.0 | No | 246.992320 | 168.413703 | !@9#%8 | 1043.315978 | 2 |
| 3 | 54 | Entrepreneur | 30689.89 | 2 | 5 | 4 | 1 | Not Specified | 0 | 6.0 | 4.0 | 632.46 | 26.544229 | 207.0 | No | 16.415452 | 81.228859 | Low_spent_Large_value_payments | 433.604773 | 1 |
| 4 | 21 | Developer | 35547.71 | 7 | 5 | 5 | 0 | NaN | 5 | NaN | 4.0 | 943.86 | 39.797764 | 368.0 | Yes | 0.000000 | 276.725394 | !@9#%8 | 288.605522 | 1 |
In [187]:
credit_df.describe()
Out[187]:
| Age | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Total_EMI_per_month | Amount_invested_monthly | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| count | 12500.000000 | 1.250000e+04 | 12500.000000 | 12500.000000 | 12500.00000 | 12500.00000 | 12500.000000 | 11657.000000 | 12264.000000 | 12500.000000 | 12500.000000 | 11387.000000 | 12500.000000 | 11935.000000 | 12366.000000 | 12500.000000 |
| mean | 113.371280 | 1.888617e+05 | 17.275120 | 21.647680 | 69.46520 | 3.43656 | 21.051440 | 31.280089 | 24.591650 | 1426.220376 | 32.291949 | 217.588127 | 1303.781040 | 643.291976 | 405.815391 | 0.871440 |
| std | 691.223297 | 1.482707e+06 | 118.518214 | 123.789969 | 455.95698 | 65.35565 | 14.859994 | 229.911798 | 183.422458 | 1155.169458 | 5.084327 | 99.638681 | 8118.261086 | 2063.324328 | 218.136964 | 0.654268 |
| min | -500.000000 | 7.005930e+03 | 0.000000 | 1.000000 | 1.00000 | -100.00000 | -5.000000 | -3.000000 | 0.000000 | 0.230000 | 20.992914 | 1.000000 | 0.000000 | 0.000000 | 0.088628 | 0.000000 |
| 25% | 24.000000 | 1.948777e+04 | 3.000000 | 4.000000 | 8.00000 | 1.00000 | 10.000000 | 9.000000 | 2.000000 | 566.072500 | 28.110034 | 141.000000 | 29.128806 | 73.810753 | 271.785749 | 0.000000 |
| 50% | 33.000000 | 3.765508e+04 | 6.000000 | 5.000000 | 13.00000 | 3.00000 | 18.000000 | 14.000000 | 4.000000 | 1166.155000 | 32.297912 | 215.000000 | 66.372879 | 134.201478 | 337.169588 | 1.000000 |
| 75% | 42.000000 | 7.289813e+04 | 7.000000 | 7.000000 | 20.00000 | 5.00000 | 28.000000 | 18.000000 | 8.000000 | 1945.962500 | 36.458660 | 298.000000 | 149.904496 | 264.555831 | 475.222487 | 1.000000 |
| max | 8592.000000 | 2.365819e+07 | 1779.000000 | 1479.000000 | 5788.00000 | 1496.00000 | 67.000000 | 4388.000000 | 2592.000000 | 4998.070000 | 49.564519 | 397.000000 | 82122.000000 | 10000.000000 | 1602.040519 | 2.000000 |
In [188]:
credit_df[credit_df['Age'] < 0]
Out[188]:
| Age | Occupation | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Type_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Payment_of_Min_Amount | Total_EMI_per_month | Amount_invested_monthly | Payment_Behaviour | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 70 | -500 | Scientist | 144546.720 | 0 | 3 | 7 | 2 | Payday Loan, and Not Specified | 15 | 7.0 | 1.0 | 1045.11 | 40.840687 | 275.0 | No | 136.988557 | 573.411590 | High_spent_Small_value_payments | 730.555853 | 1 |
| 81 | -500 | Teacher | 103353.060 | 3 | 3 | 6 | 1 | Student Loan | 4 | 9.0 | 2.0 | 1374.56 | 32.685522 | 307.0 | No | 64.001686 | NaN | High_spent_Medium_value_payments | 857.826203 | 2 |
| 122 | -500 | Journalist | 19163.220 | 3 | 7 | 7 | 2 | Personal Loan, and Payday Loan | 23 | 18.0 | 5.0 | 2226.37 | 38.080964 | 194.0 | Yes | 24.175648 | 15.206734 | High_spent_Medium_value_payments | 345.311119 | 1 |
| 165 | -500 | Doctor | 70112.780 | 1 | 7 | 3 | 2 | Mortgage Loan, and Student Loan | 5 | 3.0 | 2.0 | 877.06 | 36.976516 | 298.0 | No | 69.717948 | 433.185268 | Low_spent_Medium_value_payments | 376.069951 | 0 |
| 207 | -500 | Teacher | 94454.100 | 1 | 4 | 5 | -100 | Credit-Builder Loan, Home Equity Loan, and Hom... | 4 | 6.0 | 4.0 | 1342.61 | 26.063001 | 357.0 | NM | 152.506895 | 10000.000000 | High_spent_Medium_value_payments | 719.685063 | 2 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 11722 | -500 | Developer | 130682.940 | 2 | 331 | 1 | 4 | Payday Loan, Payday Loan, Mortgage Loan, and D... | 14 | 11.0 | 3.0 | 752.98 | 27.912234 | 344.0 | No | 263.381445 | 463.128666 | Low_spent_Large_value_payments | 604.014389 | 2 |
| 11817 | -500 | Developer | 35195.970 | 5 | 3 | 4 | 4 | Mortgage Loan, Auto Loan, Not Specified, and H... | 3 | 7.0 | 4.0 | 843.36 | 24.565802 | 338.0 | No | 93.199692 | 191.504338 | !@9#%8 | 279.095720 | 2 |
| 11839 | -500 | Architect | 8845.965 | 7 | 10 | 26 | 5 | Credit-Builder Loan, Home Equity Loan, Mortgag... | 40 | 23.0 | 6.0 | 1422.97 | 25.221897 | NaN | Yes | 33.948744 | NaN | Low_spent_Medium_value_payments | 278.220305 | 0 |
| 11917 | -500 | Teacher | 29969.660 | 8 | 6 | 18 | 1 | Auto Loan | 24 | 18.0 | 5.0 | 157.07 | 32.408860 | 266.0 | Yes | 14.380462 | 39.495791 | High_spent_Medium_value_payments | 437.970914 | 1 |
| 12053 | -500 | Doctor | 13672.200 | 7 | 5 | 13 | 1 | Auto Loan | 28 | 19.0 | 6.0 | 1419.68 | 38.426894 | 393.0 | Yes | 10.260200 | 97.019680 | Low_spent_Small_value_payments | 301.755120 | 1 |
115 rows × 20 columns
In [189]:
credit_df = credit_df[credit_df['Age'] >= 0]
In [190]:
credit_df.sort_values('Age').head(5)
Out[190]:
| Age | Occupation | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Type_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Payment_of_Min_Amount | Total_EMI_per_month | Amount_invested_monthly | Payment_Behaviour | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 3556 | 14 | Media_Manager | 7946.090 | 7 | 10 | 17 | 5 | Personal Loan, Mortgage Loan, Not Specified, P... | 54 | 27.0 | 12.0 | 1572.64 | 40.750451 | 184.0 | Yes | 30.019943 | 81.412882 | Low_spent_Small_value_payments | 266.884591 | 0 |
| 10388 | 14 | Engineer | 82941.600 | 6 | 6 | 20 | 7 | Payday Loan, Auto Loan, Auto Loan, Personal Lo... | 53 | 22.0 | 6.0 | 2598.17 | 30.264092 | 124.0 | NM | 347.228223 | 530.025784 | Low_spent_Small_value_payments | 75.325993 | 0 |
| 4215 | 14 | Engineer | 9511.795 | 6 | 7 | 20 | 7 | Student Loan, Auto Loan, Student Loan, Credit-... | 59 | 17.0 | 12.0 | 2344.43 | 34.636520 | 153.0 | NM | 28.825270 | 78.769438 | Low_spent_Small_value_payments | 250.470250 | 0 |
| 3231 | 14 | Teacher | 14015.320 | 8 | 3 | 11 | 6 | Home Equity Loan, Not Specified, Credit-Builde... | 20 | 20.0 | 7.0 | 1852.75 | 33.104663 | 118.0 | Yes | 54.001905 | 128.429618 | Low_spent_Medium_value_payments | 221.462810 | 1 |
| 10494 | 14 | Manager | 61922.240 | 6 | 9 | 18 | 9 | Payday Loan, Auto Loan, Student Loan, Credit-B... | 25 | 19.0 | 12.0 | 4678.77 | 25.661164 | 40.0 | Yes | 336.014862 | 131.642176 | High_spent_Large_value_payments | 314.161629 | 0 |
In [191]:
credit_df.sort_values('Age').tail(20)
Out[191]:
| Age | Occupation | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Type_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Payment_of_Min_Amount | Total_EMI_per_month | Amount_invested_monthly | Payment_Behaviour | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1258 | 7980 | Scientist | 44941.950 | 4 | 6 | 3 | 1 | Payday Loan | 0 | 9.0 | 2.0 | 1150.28 | 27.894834 | 367.0 | No | 22.434725 | 133.727501 | !@9#%8 | 508.454024 | 2 |
| 7445 | 8035 | Doctor | 32753.050 | 2 | 7 | 6 | 4 | Credit-Builder Loan, Debt Consolidation Loan, ... | 2 | 5.0 | 3.0 | 1300.30 | 36.662384 | 359.0 | No | 79.182771 | 63.646267 | !@9#%8 | 400.513046 | 2 |
| 10824 | 8052 | Manager | 35820.620 | 0 | 6 | 8 | 4 | Student Loan, Not Specified, Student Loan, and... | 10 | 10.0 | 3.0 | 731.26 | 26.763490 | 237.0 | No | 87.981091 | 169.322300 | Low_spent_Medium_value_payments | 298.501775 | 2 |
| 857 | 8080 | Entrepreneur | 62835.080 | 0 | 4 | 1 | 2 | Home Equity Loan, and Debt Consolidation Loan | 10 | 14.0 | 0.0 | 306.85 | 40.859818 | 392.0 | No | 67.384074 | 181.104372 | Low_spent_Medium_value_payments | 577.937220 | 2 |
| 4233 | 8081 | Accountant | 19609.380 | 9 | 7 | 23 | 8 | Payday Loan, Mortgage Loan, Home Equity Loan, ... | 16 | 23.0 | 12.0 | 4186.04 | 25.656993 | 62.0 | Yes | 87.281543 | 93.293367 | Low_spent_Small_value_payments | 276.936589 | 1 |
| 7108 | 8172 | Doctor | 23838.090 | 5 | 6 | 13 | 773 | Debt Consolidation Loan, Payday Loan, Home Equ... | 24 | 10.0 | 7.0 | 802.30 | 31.348708 | 358.0 | Yes | 53.308099 | 10000.000000 | Low_spent_Small_value_payments | 232.429877 | 1 |
| 4046 | 8179 | Scientist | 49847.550 | 9 | 1384 | 18 | 3 | Home Equity Loan, Personal Loan, and Student Loan | 21 | 18.0 | 8.0 | 2173.47 | 39.903631 | 233.0 | Yes | 71.675700 | 318.284663 | Low_spent_Medium_value_payments | 278.535887 | 0 |
| 10197 | 8198 | Journalist | 20488.290 | 8 | 5 | 33 | 9 | Auto Loan, Home Equity Loan, Home Equity Loan,... | 56 | 18.0 | 7.0 | 3218.07 | 35.075428 | 92.0 | Yes | 131.465294 | 122.316046 | Low_spent_Medium_value_payments | 207.354411 | 0 |
| 5788 | 8246 | Accountant | 42900.210 | 7 | 4 | 17 | 4 | Personal Loan, Payday Loan, Mortgage Loan, and... | 18 | 15.0 | 4.0 | 2696.09 | 31.455247 | 117.0 | Yes | 111.209311 | 150.713217 | Low_spent_Small_value_payments | 376.079222 | 0 |
| 2725 | 8249 | Developer | 35372.360 | 8 | 5 | 25 | 9 | Home Equity Loan, Auto Loan, Mortgage Loan, Au... | 15 | 24.0 | 6.0 | 3850.97 | 31.673174 | 85.0 | Yes | 173.241966 | 80.806393 | High_spent_Small_value_payments | 292.521308 | 1 |
| 11176 | 8279 | _______ | 37443.130 | 2 | 4 | 7 | 0 | NaN | 13 | 0.0 | 2.0 | 843.53 | 34.062221 | 302.0 | No | 0.000000 | NaN | High_spent_Small_value_payments | 490.572275 | 0 |
| 4738 | 8306 | Developer | 27647.080 | 7 | 6 | 13 | 5 | Not Specified, Credit-Builder Loan, Mortgage L... | 17 | 8.0 | 7.0 | 159.63 | 26.757773 | 62.0 | Yes | 100.740203 | 10000.000000 | High_spent_Small_value_payments | 294.462237 | 1 |
| 8284 | 8348 | Accountant | 19323.400 | 10 | 5 | 18 | 8 | Debt Consolidation Loan, Debt Consolidation Lo... | 48 | 15.0 | 7.0 | 3033.25 | 30.197723 | 100.0 | Yes | 99.101347 | 110.085452 | Low_spent_Medium_value_payments | 245.741535 | 0 |
| 9384 | 8403 | Developer | 7756.465 | 10 | 7 | 19 | 2 | Payday Loan, and Student Loan | 45 | 21.0 | 11.0 | 2338.09 | 25.301853 | 67.0 | Yes | 9.682065 | NaN | High_spent_Medium_value_payments | 270.333736 | 0 |
| 7630 | 8409 | Lawyer | 34413.760 | 6 | 5 | 8 | 5 | Home Equity Loan, Home Equity Loan, Home Equit... | 23 | 20.0 | 6.0 | 113.06 | 39.139526 | 69.0 | Yes | 74.328709 | 37.339760 | High_spent_Large_value_payments | 420.012864 | 1 |
| 12472 | 8425 | Writer | 18512.970 | 7 | 5 | 18 | 3 | Student Loan, Student Loan, and Mortgage Loan | 15 | 11.0 | 6.0 | 1366.56 | 29.253092 | 131.0 | Yes | 24.621173 | 63.282651 | High_spent_Medium_value_payments | 311.570927 | 1 |
| 11218 | 8481 | Journalist | 15874.010 | 1 | 3 | 5 | 0 | NaN | 12 | 6.0 | 1.0 | 1173.38 | 24.443887 | 380.0 | No | 0.000000 | 53.610174 | High_spent_Small_value_payments | 311.373243 | 0 |
| 12439 | 8490 | Lawyer | 150131.680 | 5 | 1 | 4 | 0 | NaN | 8 | -2.0 | 0.0 | 1138.36 | 30.013470 | 376.0 | No | 0.000000 | 949.847265 | Low_spent_Small_value_payments | 599.850069 | 1 |
| 9255 | 8587 | Journalist | 28286.240 | 4 | 7 | 7 | 1 | Mortgage Loan | 18 | 16.0 | 4.0 | 1406.94 | 33.599043 | 350.0 | No | 17.431672 | 41.114532 | High_spent_Medium_value_payments | 418.772462 | 1 |
| 2963 | 8592 | _______ | 81815.020 | 0 | 5 | 4 | 1 | Mortgage Loan | 4 | NaN | 3.0 | 1434.75 | 28.645438 | 394.0 | No | 53.927785 | 277.263979 | Low_spent_Medium_value_payments | 637.300069 | 2 |
In [192]:
sns.boxplot(y=credit_df['Age'])
Out[192]:
<Axes: ylabel='Age'>
In [193]:
credit_df[credit_df['Age'] >= 100].sort_values('Age')
Out[193]:
| Age | Occupation | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Type_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Payment_of_Min_Amount | Total_EMI_per_month | Amount_invested_monthly | Payment_Behaviour | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 3911 | 102 | Musician | 38460.280 | 6 | 9 | 16 | 7 | Personal Loan, Personal Loan, Home Equity Loan... | 56 | 16.0 | 11.0 | 4106.50 | 26.537577 | 73.0 | Yes | 179.135821 | 267.336505 | Low_spent_Medium_value_payments | 157.630007 | 1 |
| 2416 | 126 | Teacher | 22050.560 | 5 | 4 | 12 | 1 | Home Equity Loan | 5 | 13.0 | 4.0 | 37.42 | 28.956967 | 387.0 | NM | 13.485884 | NaN | High_spent_Medium_value_payments | 359.927004 | 1 |
| 7418 | 169 | Doctor | 50109.760 | 6 | 3 | 4 | 1 | Personal Loan | 16 | 17.0 | 0.0 | 893.62 | 27.560776 | 191.0 | NM | 22.847438 | 351.547678 | Low_spent_Large_value_payments | 331.086217 | 2 |
| 952 | 181 | _______ | 87957.020 | 2 | 5 | 9 | 4 | Home Equity Loan, Auto Loan, Credit-Builder Lo... | 9 | 1.0 | 2.0 | 811.01 | 41.470014 | 231.0 | No | 195.913703 | 265.660815 | High_spent_Medium_value_payments | 533.800649 | 0 |
| 3197 | 216 | Media_Manager | 15829.875 | 4 | 2 | 4 | 1 | Student Loan | 10 | NaN | 3.0 | 968.61 | 34.361779 | 236.0 | No | 6.719535 | NaN | Low_spent_Small_value_payments | 352.096096 | 2 |
| ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... | ... |
| 12472 | 8425 | Writer | 18512.970 | 7 | 5 | 18 | 3 | Student Loan, Student Loan, and Mortgage Loan | 15 | 11.0 | 6.0 | 1366.56 | 29.253092 | 131.0 | Yes | 24.621173 | 63.282651 | High_spent_Medium_value_payments | 311.570927 | 1 |
| 11218 | 8481 | Journalist | 15874.010 | 1 | 3 | 5 | 0 | NaN | 12 | 6.0 | 1.0 | 1173.38 | 24.443887 | 380.0 | No | 0.000000 | 53.610174 | High_spent_Small_value_payments | 311.373243 | 0 |
| 12439 | 8490 | Lawyer | 150131.680 | 5 | 1 | 4 | 0 | NaN | 8 | -2.0 | 0.0 | 1138.36 | 30.013470 | 376.0 | No | 0.000000 | 949.847265 | Low_spent_Small_value_payments | 599.850069 | 1 |
| 9255 | 8587 | Journalist | 28286.240 | 4 | 7 | 7 | 1 | Mortgage Loan | 18 | 16.0 | 4.0 | 1406.94 | 33.599043 | 350.0 | No | 17.431672 | 41.114532 | High_spent_Medium_value_payments | 418.772462 | 1 |
| 2963 | 8592 | _______ | 81815.020 | 0 | 5 | 4 | 1 | Mortgage Loan | 4 | NaN | 3.0 | 1434.75 | 28.645438 | 394.0 | No | 53.927785 | 277.263979 | Low_spent_Medium_value_payments | 637.300069 | 2 |
260 rows × 20 columns
In [194]:
credit_df = credit_df[credit_df['Age'] < 110]
In [195]:
credit_df.describe()
Out[195]:
| Age | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Total_EMI_per_month | Amount_invested_monthly | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| count | 12126.000000 | 1.212600e+04 | 12126.000000 | 12126.000000 | 12126.000000 | 12126.000000 | 12126.000000 | 11304.000000 | 11892.000000 | 12126.000000 | 12126.000000 | 11045.000000 | 12126.000000 | 11584.000000 | 11995.000000 | 12126.000000 |
| mean | 33.049398 | 1.897604e+05 | 17.275606 | 21.458601 | 68.507917 | 3.179861 | 21.053604 | 31.464084 | 24.994871 | 1426.321231 | 32.293466 | 217.614396 | 1308.972585 | 641.812701 | 406.077932 | 0.870856 |
| std | 10.810043 | 1.486221e+06 | 118.546216 | 122.910157 | 453.592030 | 64.030581 | 14.857112 | 230.020040 | 185.478101 | 1155.255348 | 5.084805 | 99.405014 | 8131.974016 | 2059.907639 | 218.484165 | 0.653729 |
| min | 14.000000 | 7.005930e+03 | 0.000000 | 1.000000 | 1.000000 | -100.000000 | -5.000000 | -3.000000 | 0.000000 | 0.230000 | 20.992914 | 1.000000 | 0.000000 | 0.000000 | 0.088628 | 0.000000 |
| 25% | 24.000000 | 1.945750e+04 | 4.000000 | 4.000000 | 8.000000 | 1.000000 | 10.000000 | 9.000000 | 2.000000 | 565.967500 | 28.118929 | 141.000000 | 29.275040 | 73.698246 | 271.752715 | 0.000000 |
| 50% | 33.000000 | 3.765508e+04 | 6.000000 | 5.000000 | 13.000000 | 3.000000 | 18.000000 | 14.000000 | 4.000000 | 1166.555000 | 32.285050 | 215.000000 | 66.196875 | 134.363758 | 337.123394 | 1.000000 |
| 75% | 41.000000 | 7.305204e+04 | 7.000000 | 7.000000 | 20.000000 | 5.000000 | 28.000000 | 18.000000 | 8.000000 | 1945.677500 | 36.452073 | 297.000000 | 149.873130 | 265.384383 | 474.903731 | 1.000000 |
| max | 102.000000 | 2.365819e+07 | 1779.000000 | 1479.000000 | 5788.000000 | 1496.000000 | 67.000000 | 4388.000000 | 2592.000000 | 4998.070000 | 49.564519 | 397.000000 | 82122.000000 | 10000.000000 | 1602.040519 | 2.000000 |
In [196]:
len(credit_df[credit_df['Num_Bank_Accounts'] > 10 ]) / len(credit_df)
Out[196]:
0.013029853207982847In [197]:
credit_df = credit_df[credit_df['Num_Bank_Accounts'] <= 10]
In [198]:
credit_df.describe()
Out[198]:
| Age | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Total_EMI_per_month | Amount_invested_monthly | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| count | 11968.000000 | 1.196800e+04 | 11968.000000 | 11968.000000 | 11968.000000 | 11968.000000 | 11968.000000 | 11155.000000 | 11739.000000 | 11968.000000 | 11968.000000 | 10902.000000 | 11968.000000 | 11435.000000 | 11840.000000 | 11968.000000 |
| mean | 33.059241 | 1.906275e+05 | 5.374749 | 21.289606 | 68.378509 | 3.202958 | 21.070438 | 31.480592 | 25.259392 | 1426.112040 | 32.290274 | 217.519171 | 1316.881794 | 637.877772 | 406.065193 | 0.870572 |
| std | 10.805276 | 1.493985e+06 | 2.589361 | 121.976512 | 454.393273 | 64.430573 | 14.868583 | 231.016220 | 186.667973 | 1155.106304 | 5.084843 | 99.394703 | 8162.702164 | 2052.261335 | 218.516737 | 0.653619 |
| min | 14.000000 | 7.005930e+03 | 0.000000 | 1.000000 | 1.000000 | -100.000000 | -5.000000 | -3.000000 | 0.000000 | 0.230000 | 20.992914 | 1.000000 | 0.000000 | 0.000000 | 0.088628 | 0.000000 |
| 25% | 24.000000 | 1.943611e+04 | 3.000000 | 4.000000 | 8.000000 | 1.000000 | 10.000000 | 9.000000 | 2.000000 | 565.175000 | 28.110034 | 141.000000 | 29.292854 | 73.446749 | 271.845069 | 0.000000 |
| 50% | 33.000000 | 3.757821e+04 | 6.000000 | 6.000000 | 13.000000 | 3.000000 | 18.000000 | 14.000000 | 4.000000 | 1167.045000 | 32.286924 | 215.000000 | 66.182932 | 133.519013 | 337.124205 | 1.000000 |
| 75% | 41.000000 | 7.305026e+04 | 7.000000 | 7.000000 | 20.000000 | 5.000000 | 28.000000 | 18.000000 | 8.000000 | 1947.157500 | 36.441411 | 297.000000 | 149.824990 | 263.272841 | 475.278504 | 1.000000 |
| max | 102.000000 | 2.365819e+07 | 10.000000 | 1479.000000 | 5788.000000 | 1496.000000 | 67.000000 | 4388.000000 | 2592.000000 | 4998.070000 | 49.564519 | 397.000000 | 82122.000000 | 10000.000000 | 1602.040519 | 2.000000 |
In [199]:
len(credit_df[credit_df['Num_Credit_Card'] > 20 ]) / len(credit_df)
Out[199]:
0.021975267379679145In [200]:
credit_df = credit_df[credit_df['Num_Credit_Card'] <= 20]
In [201]:
credit_df.describe()
Out[201]:
| Age | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Total_EMI_per_month | Amount_invested_monthly | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| count | 11705.000000 | 1.170500e+04 | 11705.000000 | 11705.000000 | 11705.000000 | 11705.000000 | 11705.000000 | 10905.000000 | 11482.000000 | 11705.000000 | 11705.000000 | 10654.000000 | 11705.000000 | 11182.000000 | 11580.000000 | 11705.000000 |
| mean | 33.051431 | 1.884028e+05 | 5.376506 | 5.539000 | 68.847074 | 3.185903 | 21.086715 | 31.781110 | 25.397579 | 1428.121067 | 32.285006 | 217.185752 | 1309.298126 | 641.162451 | 405.791718 | 0.869970 |
| std | 10.808661 | 1.483877e+06 | 2.593134 | 2.073587 | 457.588574 | 64.374527 | 14.898377 | 233.314628 | 187.626702 | 1156.429189 | 5.079989 | 99.529739 | 8140.727825 | 2058.913765 | 218.284978 | 0.653615 |
| min | 14.000000 | 7.005930e+03 | 0.000000 | 1.000000 | 1.000000 | -100.000000 | -5.000000 | -3.000000 | 0.000000 | 0.230000 | 20.992914 | 1.000000 | 0.000000 | 0.000000 | 0.088628 | 0.000000 |
| 25% | 24.000000 | 1.942803e+04 | 3.000000 | 4.000000 | 8.000000 | 1.000000 | 10.000000 | 9.000000 | 2.000000 | 565.260000 | 28.111034 | 140.000000 | 29.541637 | 73.248674 | 271.691464 | 0.000000 |
| 50% | 33.000000 | 3.764810e+04 | 6.000000 | 5.000000 | 13.000000 | 3.000000 | 18.000000 | 14.000000 | 4.000000 | 1166.910000 | 32.282544 | 215.000000 | 66.388434 | 133.865834 | 336.844798 | 1.000000 |
| 75% | 41.000000 | 7.304796e+04 | 7.000000 | 7.000000 | 20.000000 | 5.000000 | 28.000000 | 18.000000 | 8.000000 | 1952.970000 | 36.424496 | 297.000000 | 150.026842 | 263.669812 | 475.357593 | 1.000000 |
| max | 102.000000 | 2.365819e+07 | 10.000000 | 17.000000 | 5788.000000 | 1496.000000 | 67.000000 | 4388.000000 | 2592.000000 | 4998.070000 | 49.564519 | 397.000000 | 82122.000000 | 10000.000000 | 1602.040519 | 2.000000 |
In [202]:
credit_df = credit_df[credit_df['Interest_Rate'] <= 40]
In [203]:
credit_df.describe()
Out[203]:
| Age | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Total_EMI_per_month | Amount_invested_monthly | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| count | 11489.000000 | 1.148900e+04 | 11489.000000 | 11489.000000 | 11489.000000 | 11489.000000 | 11489.000000 | 10697.000000 | 11272.000000 | 11489.000000 | 11489.000000 | 10456.000000 | 11489.000000 | 10976.000000 | 11365.000000 | 11489.000000 |
| mean | 33.051005 | 1.866428e+05 | 5.379058 | 5.539473 | 14.581774 | 3.085299 | 21.071111 | 32.144433 | 25.610628 | 1428.023787 | 32.274542 | 217.205337 | 1309.664252 | 643.204912 | 405.469860 | 0.869005 |
| std | 10.806912 | 1.473204e+06 | 2.592378 | 2.073895 | 8.769424 | 63.606989 | 14.867793 | 235.555903 | 188.942465 | 1156.841447 | 5.081960 | 99.622916 | 8147.400974 | 2063.638237 | 217.870485 | 0.652951 |
| min | 14.000000 | 7.005930e+03 | 0.000000 | 1.000000 | 1.000000 | -100.000000 | -5.000000 | -3.000000 | 0.000000 | 0.230000 | 20.992914 | 1.000000 | 0.000000 | 0.000000 | 0.088628 | 0.000000 |
| 25% | 24.000000 | 1.939932e+04 | 3.000000 | 4.000000 | 7.000000 | 1.000000 | 10.000000 | 9.000000 | 2.000000 | 565.280000 | 28.109667 | 140.000000 | 29.579944 | 73.167755 | 271.696442 | 0.000000 |
| 50% | 33.000000 | 3.747152e+04 | 6.000000 | 5.000000 | 13.000000 | 3.000000 | 18.000000 | 14.000000 | 4.000000 | 1166.470000 | 32.257865 | 215.000000 | 66.140607 | 133.820509 | 336.826381 | 1.000000 |
| 75% | 41.000000 | 7.288608e+04 | 7.000000 | 7.000000 | 20.000000 | 5.000000 | 28.000000 | 18.000000 | 8.000000 | 1950.620000 | 36.403191 | 297.000000 | 149.469529 | 263.650717 | 474.571468 | 1.000000 |
| max | 102.000000 | 2.365819e+07 | 10.000000 | 17.000000 | 34.000000 | 1496.000000 | 67.000000 | 4388.000000 | 2592.000000 | 4998.070000 | 49.564519 | 397.000000 | 82122.000000 | 10000.000000 | 1602.040519 | 2.000000 |
In [204]:
len(credit_df[credit_df['Num_of_Loan'] > 20 ])
Out[204]:
60In [205]:
credit_df = credit_df[(credit_df['Num_of_Loan'] <= 20) & (credit_df['Num_of_Loan'] >= 0)]
In [206]:
credit_df.describe()
Out[206]:
| Age | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Total_EMI_per_month | Amount_invested_monthly | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| count | 10962.000000 | 1.096200e+04 | 10962.000000 | 10962.000000 | 10962.000000 | 10962.000000 | 10962.000000 | 10211.000000 | 10756.000000 | 10962.000000 | 10962.000000 | 9967.000000 | 10962.000000 | 10468.000000 | 10840.000000 | 10962.000000 |
| mean | 33.044426 | 1.883743e+05 | 5.386882 | 5.538405 | 14.592501 | 3.552728 | 21.079091 | 32.153168 | 25.391595 | 1430.842134 | 32.280912 | 216.911006 | 1322.954566 | 646.217707 | 405.524273 | 0.867907 |
| std | 10.814471 | 1.479179e+06 | 2.592619 | 2.073381 | 8.771167 | 2.449676 | 14.865527 | 235.514645 | 187.830090 | 1157.869902 | 5.083798 | 99.455191 | 8210.229115 | 2068.395275 | 218.131576 | 0.651802 |
| min | 14.000000 | 7.005930e+03 | 0.000000 | 1.000000 | 1.000000 | 0.000000 | -5.000000 | -3.000000 | 0.000000 | 0.230000 | 20.992914 | 1.000000 | 0.000000 | 0.000000 | 0.088628 | 0.000000 |
| 25% | 24.000000 | 1.940034e+04 | 3.000000 | 4.000000 | 7.000000 | 2.000000 | 10.000000 | 9.000000 | 2.000000 | 566.242500 | 28.134649 | 140.000000 | 29.617499 | 73.265262 | 271.841052 | 0.000000 |
| 50% | 33.000000 | 3.768010e+04 | 6.000000 | 5.000000 | 13.000000 | 3.000000 | 18.000000 | 14.000000 | 4.000000 | 1166.420000 | 32.242901 | 214.000000 | 66.333116 | 133.543758 | 336.830727 | 1.000000 |
| 75% | 41.000000 | 7.305204e+04 | 7.000000 | 7.000000 | 20.000000 | 5.000000 | 28.000000 | 18.000000 | 8.000000 | 1961.112500 | 36.409833 | 296.000000 | 150.112234 | 265.694698 | 474.525972 | 1.000000 |
| max | 102.000000 | 2.365819e+07 | 10.000000 | 17.000000 | 34.000000 | 19.000000 | 67.000000 | 4388.000000 | 2592.000000 | 4998.070000 | 49.564519 | 397.000000 | 82122.000000 | 10000.000000 | 1602.040519 | 2.000000 |
In [207]:
credit_df = credit_df[credit_df['Delay_from_due_date'] >= 0]
In [208]:
len(credit_df[credit_df['Num_of_Delayed_Payment'] > 30])
Out[208]:
80In [209]:
credit_df = credit_df[(credit_df['Num_of_Delayed_Payment'] <= 30) & (credit_df['Num_of_Delayed_Payment'] >= 0)]
In [210]:
credit_df.info()
<class 'pandas.core.frame.DataFrame'>
Index: 10005 entries, 0 to 12498
Data columns (total 20 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Age 10005 non-null int64
1 Occupation 10005 non-null object
2 Annual_Income 10005 non-null float64
3 Num_Bank_Accounts 10005 non-null int64
4 Num_Credit_Card 10005 non-null int64
5 Interest_Rate 10005 non-null int64
6 Num_of_Loan 10005 non-null int64
7 Type_of_Loan 8894 non-null object
8 Delay_from_due_date 10005 non-null int64
9 Num_of_Delayed_Payment 10005 non-null float64
10 Num_Credit_Inquiries 9817 non-null float64
11 Outstanding_Debt 10005 non-null float64
12 Credit_Utilization_Ratio 10005 non-null float64
13 Credit_History_Age 9107 non-null float64
14 Payment_of_Min_Amount 10005 non-null object
15 Total_EMI_per_month 10005 non-null float64
16 Amount_invested_monthly 9550 non-null float64
17 Payment_Behaviour 10005 non-null object
18 Monthly_Balance 9896 non-null float64
19 Credit_Score 10005 non-null int64
dtypes: float64(9), int64(7), object(4)
memory usage: 1.6+ MB
In [211]:
credit_df['Num_Credit_Inquiries'] = credit_df['Num_Credit_Inquiries'].fillna(0)
<ipython-input-211-17ca6241ab57>:1: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead
See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy
credit_df['Num_Credit_Inquiries'] = credit_df['Num_Credit_Inquiries'].fillna(0)
In [212]:
credit_df.isna().mean()
Out[212]:
Age 0.000000
Occupation 0.000000
Annual_Income 0.000000
Num_Bank_Accounts 0.000000
Num_Credit_Card 0.000000
Interest_Rate 0.000000
Num_of_Loan 0.000000
Type_of_Loan 0.111044
Delay_from_due_date 0.000000
Num_of_Delayed_Payment 0.000000
Num_Credit_Inquiries 0.000000
Outstanding_Debt 0.000000
Credit_Utilization_Ratio 0.000000
Credit_History_Age 0.089755
Payment_of_Min_Amount 0.000000
Total_EMI_per_month 0.000000
Amount_invested_monthly 0.045477
Payment_Behaviour 0.000000
Monthly_Balance 0.010895
Credit_Score 0.000000
dtype: float64In [213]:
sns.displot(credit_df['Credit_History_Age'])
Out[213]:
<seaborn.axisgrid.FacetGrid at 0x79ae6435ac80>
In [214]:
sns.displot(credit_df['Amount_invested_monthly'])
Out[214]:
<seaborn.axisgrid.FacetGrid at 0x79ae643591b0>
In [215]:
sns.displot(credit_df['Monthly_Balance'])
Out[215]:
<seaborn.axisgrid.FacetGrid at 0x79ae64024340>
In [216]:
credit_df = credit_df.fillna(credit_df.median(numeric_only=True))
In [217]:
credit_df.isna().mean()
Out[217]:
Age 0.000000
Occupation 0.000000
Annual_Income 0.000000
Num_Bank_Accounts 0.000000
Num_Credit_Card 0.000000
Interest_Rate 0.000000
Num_of_Loan 0.000000
Type_of_Loan 0.111044
Delay_from_due_date 0.000000
Num_of_Delayed_Payment 0.000000
Num_Credit_Inquiries 0.000000
Outstanding_Debt 0.000000
Credit_Utilization_Ratio 0.000000
Credit_History_Age 0.000000
Payment_of_Min_Amount 0.000000
Total_EMI_per_month 0.000000
Amount_invested_monthly 0.000000
Payment_Behaviour 0.000000
Monthly_Balance 0.000000
Credit_Score 0.000000
dtype: float64In [218]:
credit_df.head()
Out[218]:
| Age | Occupation | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Type_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | Credit_Utilization_Ratio | Credit_History_Age | Payment_of_Min_Amount | Total_EMI_per_month | Amount_invested_monthly | Payment_Behaviour | Monthly_Balance | Credit_Score | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 23 | Scientist | 19114.12 | 3 | 4 | 3 | 4 | Auto Loan, Credit-Builder Loan, Personal Loan,... | 3 | 7.0 | 4.0 | 809.98 | 26.822620 | 265.0 | No | 49.574949 | 80.415295 | High_spent_Small_value_payments | 312.494089 | 2 |
| 1 | 28 | _______ | 34847.84 | 2 | 4 | 6 | 1 | Credit-Builder Loan | 3 | 4.0 | 2.0 | 605.03 | 24.464031 | 319.0 | No | 18.816215 | 104.291825 | Low_spent_Small_value_payments | 470.690627 | 1 |
| 2 | 34 | _______ | 143162.64 | 1 | 5 | 8 | 3 | Auto Loan, Auto Loan, and Not Specified | 5 | 8.0 | 3.0 | 1303.01 | 28.616735 | 213.0 | No | 246.992320 | 168.413703 | !@9#%8 | 1043.315978 | 2 |
| 3 | 54 | Entrepreneur | 30689.89 | 2 | 5 | 4 | 1 | Not Specified | 0 | 6.0 | 4.0 | 632.46 | 26.544229 | 207.0 | No | 16.415452 | 81.228859 | Low_spent_Large_value_payments | 433.604773 | 1 |
| 6 | 33 | Lawyer | 131313.40 | 0 | 1 | 8 | 2 | Credit-Builder Loan, and Mortgage Loan | 0 | 3.0 | 2.0 | 352.16 | 32.200509 | 367.0 | NM | 137.644605 | 378.171253 | High_spent_Medium_value_payments | 858.462474 | 2 |
In [219]:
# 문제
# Type_of_Loan 의 모든 대출 상품을 변수에 저장
# NaN인 데이터는 'No Loan' 으로 대체
# 대출상품 만큼의 컬럼을 만들고 해당 대출 상품을 받았다면 1 아니면 0으로 데이터 처리
credit_df['Type_of_Loan'] = credit_df['Type_of_Loan'].str.replace('and ', '')
In [222]:
credit_df.isna().mean()
Out[222]:
Age 0.000000
Occupation 0.000000
Annual_Income 0.000000
Num_Bank_Accounts 0.000000
Num_Credit_Card 0.000000
Interest_Rate 0.000000
Num_of_Loan 0.000000
Type_of_Loan 0.111044
Delay_from_due_date 0.000000
Num_of_Delayed_Payment 0.000000
Num_Credit_Inquiries 0.000000
Outstanding_Debt 0.000000
Credit_Utilization_Ratio 0.000000
Credit_History_Age 0.000000
Payment_of_Min_Amount 0.000000
Total_EMI_per_month 0.000000
Amount_invested_monthly 0.000000
Payment_Behaviour 0.000000
Monthly_Balance 0.000000
Credit_Score 0.000000
dtype: float64In [223]:
credit_df['Type_of_Loan'] = credit_df['Type_of_Loan'].fillna('No Loan')
In [224]:
credit_df.isna().mean()
Out[224]:
Age 0.0
Occupation 0.0
Annual_Income 0.0
Num_Bank_Accounts 0.0
Num_Credit_Card 0.0
Interest_Rate 0.0
Num_of_Loan 0.0
Type_of_Loan 0.0
Delay_from_due_date 0.0
Num_of_Delayed_Payment 0.0
Num_Credit_Inquiries 0.0
Outstanding_Debt 0.0
Credit_Utilization_Ratio 0.0
Credit_History_Age 0.0
Payment_of_Min_Amount 0.0
Total_EMI_per_month 0.0
Amount_invested_monthly 0.0
Payment_Behaviour 0.0
Monthly_Balance 0.0
Credit_Score 0.0
dtype: float64In [226]:
type_list = set(credit_df['Type_of_Loan'].str.split(', ').sum())
type_list
Out[226]:
{'Auto Loan',
'Credit-Builder Loan',
'Debt Consolidation Loan',
'Home Equity Loan',
'Mortgage Loan',
'No Loan',
'Not Specified',
'Payday Loan',
'Personal Loan',
'Student Loan'}In [227]:
for i in type_list:
credit_df[i] = credit_df['Type_of_Loan'].apply(lambda x: 1 if i in x else 0)
In [228]:
credit_df.head()
Out[228]:
| Age | Occupation | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Type_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | ... | Home Equity Loan | No Loan | Auto Loan | Personal Loan | Not Specified | Credit-Builder Loan | Debt Consolidation Loan | Payday Loan | Mortgage Loan | Student Loan | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 23 | Scientist | 19114.12 | 3 | 4 | 3 | 4 | Auto Loan, Credit-Builder Loan, Personal Loan,... | 3 | 7.0 | ... | 1 | 0 | 1 | 1 | 0 | 1 | 0 | 0 | 0 | 0 |
| 1 | 28 | _______ | 34847.84 | 2 | 4 | 6 | 1 | Credit-Builder Loan | 3 | 4.0 | ... | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 |
| 2 | 34 | _______ | 143162.64 | 1 | 5 | 8 | 3 | Auto Loan, Auto Loan, Not Specified | 5 | 8.0 | ... | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 0 | 0 |
| 3 | 54 | Entrepreneur | 30689.89 | 2 | 5 | 4 | 1 | Not Specified | 0 | 6.0 | ... | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 0 |
| 6 | 33 | Lawyer | 131313.40 | 0 | 1 | 8 | 2 | Credit-Builder Loan, Mortgage Loan | 0 | 3.0 | ... | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 1 | 0 |
5 rows × 30 columns
In [229]:
credit_df.drop('Type_of_Loan', axis=1, inplace=True)
In [230]:
credit_df.info()
<class 'pandas.core.frame.DataFrame'>
Index: 10005 entries, 0 to 12498
Data columns (total 29 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Age 10005 non-null int64
1 Occupation 10005 non-null object
2 Annual_Income 10005 non-null float64
3 Num_Bank_Accounts 10005 non-null int64
4 Num_Credit_Card 10005 non-null int64
5 Interest_Rate 10005 non-null int64
6 Num_of_Loan 10005 non-null int64
7 Delay_from_due_date 10005 non-null int64
8 Num_of_Delayed_Payment 10005 non-null float64
9 Num_Credit_Inquiries 10005 non-null float64
10 Outstanding_Debt 10005 non-null float64
11 Credit_Utilization_Ratio 10005 non-null float64
12 Credit_History_Age 10005 non-null float64
13 Payment_of_Min_Amount 10005 non-null object
14 Total_EMI_per_month 10005 non-null float64
15 Amount_invested_monthly 10005 non-null float64
16 Payment_Behaviour 10005 non-null object
17 Monthly_Balance 10005 non-null float64
18 Credit_Score 10005 non-null int64
19 Home Equity Loan 10005 non-null int64
20 No Loan 10005 non-null int64
21 Auto Loan 10005 non-null int64
22 Personal Loan 10005 non-null int64
23 Not Specified 10005 non-null int64
24 Credit-Builder Loan 10005 non-null int64
25 Debt Consolidation Loan 10005 non-null int64
26 Payday Loan 10005 non-null int64
27 Mortgage Loan 10005 non-null int64
28 Student Loan 10005 non-null int64
dtypes: float64(9), int64(17), object(3)
memory usage: 2.3+ MB
In [241]:
# Occupation
# '-----' 를 'Unknown'
credit_df['Occupation'] = credit_df['Occupation'].replace('_______', 'Unknown')
credit_df['Occupation'].unique()
Out[241]:
array(['Scientist', 'Unknown', 'Entrepreneur', 'Lawyer', 'Journalist',
'Teacher', 'Manager', 'Accountant', 'Musician', 'Mechanic',
'Writer', 'Architect', 'Engineer', 'Developer', 'Media_Manager',
'Doctor'], dtype=object)In [244]:
# Payment_of_Min_Amount
# '!@9#%8' 를 'Unknown'
credit_df['Payment_of_Min_Amount'] = credit_df['Payment_of_Min_Amount'].replace('!@9#%8', 'Unknown')
credit_df['Payment_of_Min_Amount'].unique()
Out[244]:
array(['No', 'NM', 'Yes'], dtype=object)In [246]:
# Payment_Behaviour
credit_df['Payment_Behaviour'] = credit_df['Payment_Behaviour'].replace('!@9#%8', 'Unknown')
credit_df['Payment_Behaviour'].unique()
Out[246]:
array(['High_spent_Small_value_payments',
'Low_spent_Small_value_payments', 'Unknown',
'Low_spent_Large_value_payments',
'High_spent_Medium_value_payments',
'Low_spent_Medium_value_payments',
'High_spent_Large_value_payments'], dtype=object)In [247]:
# 위 object를 원핫인코딩
credit_df = pd.get_dummies(credit_df, columns=['Occupation', 'Payment_of_Min_Amount', 'Payment_Behaviour'])
credit_df.head()
Out[247]:
| Age | Annual_Income | Num_Bank_Accounts | Num_Credit_Card | Interest_Rate | Num_of_Loan | Delay_from_due_date | Num_of_Delayed_Payment | Num_Credit_Inquiries | Outstanding_Debt | ... | Payment_of_Min_Amount_NM | Payment_of_Min_Amount_No | Payment_of_Min_Amount_Yes | Payment_Behaviour_High_spent_Large_value_payments | Payment_Behaviour_High_spent_Medium_value_payments | Payment_Behaviour_High_spent_Small_value_payments | Payment_Behaviour_Low_spent_Large_value_payments | Payment_Behaviour_Low_spent_Medium_value_payments | Payment_Behaviour_Low_spent_Small_value_payments | Payment_Behaviour_Unknown | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | 23 | 19114.12 | 3 | 4 | 3 | 4 | 3 | 7.0 | 4.0 | 809.98 | ... | False | True | False | False | False | True | False | False | False | False |
| 1 | 28 | 34847.84 | 2 | 4 | 6 | 1 | 3 | 4.0 | 2.0 | 605.03 | ... | False | True | False | False | False | False | False | False | True | False |
| 2 | 34 | 143162.64 | 1 | 5 | 8 | 3 | 5 | 8.0 | 3.0 | 1303.01 | ... | False | True | False | False | False | False | False | False | False | True |
| 3 | 54 | 30689.89 | 2 | 5 | 4 | 1 | 0 | 6.0 | 4.0 | 632.46 | ... | False | True | False | False | False | False | True | False | False | False |
| 6 | 33 | 131313.40 | 0 | 1 | 8 | 2 | 0 | 3.0 | 2.0 | 352.16 | ... | True | False | False | False | True | False | False | False | False | False |
5 rows × 52 columns
In [248]:
credit_df.info()
<class 'pandas.core.frame.DataFrame'>
Index: 10005 entries, 0 to 12498
Data columns (total 52 columns):
# Column Non-Null Count Dtype
--- ------ -------------- -----
0 Age 10005 non-null int64
1 Annual_Income 10005 non-null float64
2 Num_Bank_Accounts 10005 non-null int64
3 Num_Credit_Card 10005 non-null int64
4 Interest_Rate 10005 non-null int64
5 Num_of_Loan 10005 non-null int64
6 Delay_from_due_date 10005 non-null int64
7 Num_of_Delayed_Payment 10005 non-null float64
8 Num_Credit_Inquiries 10005 non-null float64
9 Outstanding_Debt 10005 non-null float64
10 Credit_Utilization_Ratio 10005 non-null float64
11 Credit_History_Age 10005 non-null float64
12 Total_EMI_per_month 10005 non-null float64
13 Amount_invested_monthly 10005 non-null float64
14 Monthly_Balance 10005 non-null float64
15 Credit_Score 10005 non-null int64
16 Home Equity Loan 10005 non-null int64
17 No Loan 10005 non-null int64
18 Auto Loan 10005 non-null int64
19 Personal Loan 10005 non-null int64
20 Not Specified 10005 non-null int64
21 Credit-Builder Loan 10005 non-null int64
22 Debt Consolidation Loan 10005 non-null int64
23 Payday Loan 10005 non-null int64
24 Mortgage Loan 10005 non-null int64
25 Student Loan 10005 non-null int64
26 Occupation_Accountant 10005 non-null bool
27 Occupation_Architect 10005 non-null bool
28 Occupation_Developer 10005 non-null bool
29 Occupation_Doctor 10005 non-null bool
30 Occupation_Engineer 10005 non-null bool
31 Occupation_Entrepreneur 10005 non-null bool
32 Occupation_Journalist 10005 non-null bool
33 Occupation_Lawyer 10005 non-null bool
34 Occupation_Manager 10005 non-null bool
35 Occupation_Mechanic 10005 non-null bool
36 Occupation_Media_Manager 10005 non-null bool
37 Occupation_Musician 10005 non-null bool
38 Occupation_Scientist 10005 non-null bool
39 Occupation_Teacher 10005 non-null bool
40 Occupation_Unknown 10005 non-null bool
41 Occupation_Writer 10005 non-null bool
42 Payment_of_Min_Amount_NM 10005 non-null bool
43 Payment_of_Min_Amount_No 10005 non-null bool
44 Payment_of_Min_Amount_Yes 10005 non-null bool
45 Payment_Behaviour_High_spent_Large_value_payments 10005 non-null bool
46 Payment_Behaviour_High_spent_Medium_value_payments 10005 non-null bool
47 Payment_Behaviour_High_spent_Small_value_payments 10005 non-null bool
48 Payment_Behaviour_Low_spent_Large_value_payments 10005 non-null bool
49 Payment_Behaviour_Low_spent_Medium_value_payments 10005 non-null bool
50 Payment_Behaviour_Low_spent_Small_value_payments 10005 non-null bool
51 Payment_Behaviour_Unknown 10005 non-null bool
dtypes: bool(26), float64(9), int64(17)
memory usage: 2.3 MB
In [254]:
from sklearn.model_selection import train_test_split
In [255]:
X_train, X_test, y_train, y_test = train_test_split(credit_df.drop('Credit_Score', axis=1), credit_df['Credit_Score'], test_size=0.2, random_state=2024)
In [256]:
X_train.shape, X_test.shape
Out[256]:
((8004, 51), (2001, 51))In [257]:
y_train.shape, y_test.shape
Out[257]:
((8004,), (2001,))2. lightGBM(LGBM)¶
- Microsoft에서 개발한 Gradient Boosting Framework
- 리프 중심 히스토그램 기반 알고리즘
- 작은 데이터셋에서도 높은 성능을 보이며, 특히 대용량 데이터셋에서 다른 알고리즘보다 빠르게 학습
- 메모리 사용량이 상대적으로 적은편
- 적은 데이터셋을 사용할 경우 과적합 가능성이 매우 큼(일반적으로 데이터가 10,000개 이상은 사용해야함)
- 조기 중단(early stopping)을 지원
2-1. 리프 중심 히스토그램 기반 알고리즘¶
- 트리를 균형적으로 분할하는 것이 아니라, 최대한 불균형하게 분할
- 특성들의 분포를 히스토그램으로 나타내고, 해당 히스토그램을 이용하여 빠르게 후보 분할 기준을 선택
- 후보 분할 기준 중에서 최적의 분할 기준을 선택하기 위해, 데이터 포인트들을 히스토그램에 올바르게 배치하고 이를 이용하여 최적의 분할 기준을 선택
2-2 GBM(Gradient Boosting Model)¶
- 순차적으로 모델을 학습시킴
- 첫 번째 모델을 학습시키고, 두 번째 모델은 첫 번째 모델의 오류를 학습하는 식으로 진행(이런 방식으로 각 모델이 이전 모델의 오류를 보완)
- 부스팅에서는 각 데이터 포인트에 가중치를 부여하지만, 이후 모델이 학습되면서 잘못 예측된 데이터 포인트의 가중치를 증가시켜 다음 모델이 디 데이터 포인트에 더 주의를 기울이도록 함
- 트리가 모두 학습된 후 예측 결과를 결합하여 최종 예측을 만드는데 일반적으로 분류 문데에서는 다수결 투표 방식으로, 회귀 문제에서는 예측값의 평균을 사용
2-3. 부스팅 모델의 주요 개념¶
- 약한 학습기(Weak Learner): 단독으로는 성능이 좋지 않은 간단한 모델(주로 깊이가 얕은 결정 트리, 깊이가 1인 매우 간단한 모델)
- 약한 학습기를 순차적으로 학습시키고 그 다음에는 첫 번째 학습기의 오류를 보완하는 두 번째 학습기를 학습시킴
In [258]:
from lightgbm import LGBMClassifier
In [260]:
base_model = LGBMClassifier(random_state=2024)
In [261]:
base_model.fit(X_train, y_train)
[LightGBM] [Warning] Found whitespace in feature_names, replace with underlines
[LightGBM] [Info] Auto-choosing row-wise multi-threading, the overhead of testing was 0.004698 seconds.
You can set `force_row_wise=true` to remove the overhead.
And if memory is not enough, you can set `force_col_wise=true`.
[LightGBM] [Info] Total Bins 2103
[LightGBM] [Info] Number of data points in the train set: 8004, number of used features: 51
[LightGBM] [Info] Start training from score -1.246598
[LightGBM] [Info] Start training from score -0.576753
[LightGBM] [Info] Start training from score -1.891803
Out[261]:
LGBMClassifier(random_state=2024)On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
LGBMClassifier(random_state=2024)In [262]:
pred = base_model.predict(X_test)
In [263]:
from sklearn.metrics import accuracy_score, confusion_matrix, classification_report, roc_auc_score
In [264]:
accuracy_score(y_test, pred)
Out[264]:
0.7386306846576711In [265]:
confusion_matrix(y_test, pred)
Out[265]:
array([[442, 140, 35],
[131, 858, 99],
[ 10, 108, 178]])In [268]:
print(classification_report(y_test, pred))
precision recall f1-score support
0 0.76 0.72 0.74 617
1 0.78 0.79 0.78 1088
2 0.57 0.60 0.59 296
accuracy 0.74 2001
macro avg 0.70 0.70 0.70 2001
weighted avg 0.74 0.74 0.74 2001
In [266]:
proba = base_model.predict_proba(X_test)
proba
Out[266]:
array([[5.31741083e-02, 5.47962223e-01, 3.98863668e-01],
[3.10313153e-03, 9.96804047e-01, 9.28219692e-05],
[9.62549410e-01, 3.73701069e-02, 8.04835753e-05],
...,
[8.47182625e-01, 1.52551839e-01, 2.65535858e-04],
[1.93275922e-01, 8.06637389e-01, 8.66884417e-05],
[9.69702722e-03, 1.60090661e-01, 8.30212312e-01]])In [269]:
5.31741083e-02, 5.47962223e-01, 3.98863668e-01
Out[269]:
(0.0531741083, 0.547962223, 0.398863668)In [270]:
roc_auc_score(y_test, proba, multi_class='ovr')
Out[270]:
0.8943814823196526In [ ]: