Skip to content

Philip-Cho/BitaminProject_Loan-Classification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

27 Commits
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

BITAminπŸ’Š Review Project (Team No.2)

TopicπŸ’΅ : Loan Approval Prediction(λŒ€μΆœ 승인 μ—¬λΆ€ 예츑)

Coworkers 🎈

  • βœ” 김닀희 @DieKim
  • βœ” 이쀀석 @timointhebush
  • βœ” 쑰성택 @Philip-Cho

Data βš’

  • Loan Application Data
  • https://www.kaggle.com/vipin20/loan-application-data
  • Predict whether loan will be approved by using the informations from loan application form
  • λŒ€μΆœ μ‹ μ²­μ„œμ˜ λ‚΄μš©μ„ λ°”νƒ•μœΌλ‘œ λŒ€μΆœ 적격 심사 예츑

Project Period πŸ“„

  • 2021/07/12 ~ 2021/07/30

What we can Think of πŸ€”

As a result of our Analysis, Credit_History, LoanAmount and Income related variables were most important features for Loan Approval
μš°λ¦¬κ°€ μ£Όμš” λͺ¨λΈλ‘œ μ‚¬μš©ν•œ 랜덀포레슀트 λͺ¨λΈμ— λ”°λ₯΄λ©΄, μ‹ μš©κΈ°λ‘κ³Ό λŒ€μΆœκ·œλͺ¨ 및 μ†Œλ“ κ΄€λ ¨ λ³€μˆ˜λ“€μ΄ λŒ€μΆœμŠΉμΈμ— μ€‘μš”ν•œ μš”μ†Œλ“€μ΄ λ˜μ—ˆλ‹€.

  1. Credit_History (μ‹ μš© 기둝)
    People with Credit History were more likely to get a loan approved. The financial industry (except BigTech/FinTech) evaluates an individual's credit ratings based on their credit history, such as whether they have used a credit card or not, and past loan experience. However, this method has a disadvantage in that it cannot accurately evaluate thin-filers(people who don't have credit history but ability to repay the debt - ex. student, housewife etc). We can see these characteristics from the data we used. In this situation, we need to think about a more reasonable way to rate individual's credit. Many of Fintech companies are making these effort.
  • μ‹ μš© 기둝이 μžˆλŠ” μ‚¬λžŒμ΄ λŒ€μΆœμ„ 승인 받을 ν™•λ₯ μ΄ 더 λ†’κ²Œ λ‚˜νƒ€λ‚¬λ‹€. 기쑴의 κΈˆμœ΅μ—…κ³„λŠ” μ‹ μš©μΉ΄λ“œ μ‚¬μš© μ—¬λΆ€, κ³Όκ±° λŒ€μΆœ κ²½ν—˜ λ“±μ˜ μ‹ μš© 기둝듀을 ν† λŒ€λ‘œ 개인의 μ‹ μš©λ“±κΈ‰μ„ ν‰κ°€ν•œλ‹€. ν•œνŽΈ, μ΄λŸ¬ν•œ 방식은 μ‹€μ§ˆμ μΈ μ±„λ¬΄μƒν™˜ λŠ₯λ ₯은 μžˆμœΌλ‚˜ μ‹ μš© 기둝(금육거래 기둝)이 μ—†λŠ” μ‹ -파일러(Thin Filer)듀에 λŒ€ν•œ μ •ν™•ν•œ 평가λ₯Ό ν•  수 μ—†λ‹€λŠ” 단점이 μžˆλ‹€. 뢄석 결과에 λ”°λ₯΄λ©΄, ν•΄λ‹Ή λ°μ΄ν„°μ—λŠ” μ΄λŸ¬ν•œ νŠΉμ„±λ“€μ΄ λ‚˜νƒ€λ‚˜κ³  μžˆλ‹€. λ”°λΌμ„œ, μš°λ¦¬λŠ” μ΄λŸ¬ν•œ μ‹ -νŒŒμΌλŸ¬λ“€μ„ μœ„ν•œ λ°”λžŒμ§ν•œ μ‹ μš©ν‰κ°€ 방식을 고민해봐야 ν•œλ‹€. 그리고 ν˜„μž¬ ν•€ν…Œν¬ μ‹œμž₯μ—μ„œμ˜ μžμ‚°κ΄€λ¦¬ 기업듀은 μ΄λŸ¬ν•œ λ…Έλ ₯을 ν•˜κ³  μžˆλ‹€.
  1. LoanAmount and Income (λŒ€μΆœκΈˆμ•‘κ³Ό μ†Œλ“κ³Όμ˜ 관계)
    Data such as occupation and income level are important data that can be used to evaluate whether a loan applicant has a ability to repay the loan. According to analysis of the project, low-income group can only get small loans, and the higher their income level, the more likely they are to be approved for a large loan. Here we need to think about how to efficiently provide loans to low-income earners who need high-value loans for the purpose of mortgages, etc. However, we also have to consider what happened at 2008, 'Subprime Mortgage Loan Crisis'.
  • 직업, μ†Œλ“μˆ˜μ€€ λ“±μ˜ μžλ£Œλ“€μ€ λŒ€μΆœ μ‹ μ²­μžκ°€ λŒ€μΆœκΈˆμ„ μƒν™˜ν•  λŠ₯λ ₯이 μžˆλŠ”μ§€ 평가할 수 μžˆλŠ” μ€‘μš”ν•œ 데이터이닀. 이 ν”„λ‘œμ νŠΈμ˜ 뢄석에 λ”°λ₯΄λ©΄, μ €μ†Œλ“μžλŠ” μ†Œμ•‘ λŒ€μΆœλ§Œ 받을 수 있으며 μ†Œλ“μˆ˜μ€€μ΄ λ†’μ•„μ§ˆμˆ˜λ‘ κ³ μ•‘ λŒ€μΆœμ„ μŠΉμΈλ°›μ„ ν™•λ₯ μ΄ λ†’μ•„μ§„λ‹€. μ—¬κΈ°μ„œ μš°λ¦¬λŠ” μ €μ†Œλ“μž μ€‘μ—μ„œ μ£Όνƒλ‹΄λ³΄λŒ€μΆœ λ“±μ˜ λͺ©μ μœΌλ‘œ κ³ μ•‘λŒ€μΆœμ΄ ν•„μš”ν•œ μ‚¬λžŒλ“€μ—κ²Œ μ–΄λ–»κ²Œ ν•˜λ©΄ 효율적으둜 λŒ€μΆœμ΄ κ°€λŠ₯할지에 λŒ€ν•΄ 고민이 ν•„μš”ν•˜λ‹€. ν•œνŽΈ, κ³Όκ±° 2008λ…„ λ°œμƒν–ˆλ˜ 'μ„œλΈŒν”„λΌμž„ λͺ¨κΈ°μ§€λ‘  μ‚¬νƒœ' λ“±μ˜ 사둀에 λŒ€ν•΄μ„œλ„ κ³ λ €ν•΄μ•Ό ν•œλ‹€.

About

This project was done by Bitamin Club Members. Aim of this project is to predict whether loan approved or not.

Resources

Stars

0 stars

Watchers

1 watching

Forks

Releases

No releases published

Packages

 
 
 

Contributors