Banks are essentials bodies of the societies which not only serve as a secret vault for storing money but also help people in time of need by providing loans and credit. However, the rate of loan default is increasing exponentially. According to Forbes, say that nearly 40% of borrowers are expected to default on their stiudent loans by 2023. Considering the statistics, instead of making money from the loan interest, bank will suffer a huge capital loss. In order to prevent the loss, it is very important to have a system in place which can accurately predict the loan defaulters even before approving the loan.
Designing a machine learning model to predict defaulter based on the basis of the historic customer behaviour.
- A Financial organization wants to predict the possible defaulters for the consumer loans product.
- They have data about historic customer behavior based on what they have observed.
- Hence when they acquire new customers they want to predict who is riskier and who is not.
The Dataset is a synthetic collection of 2,52,000 loan applicant information over 13 features/fields. It consists of fields that describes basic demographics like Income, Age, Experience, Profession, Martial Status, House-ownership, car-ownership, current job years, current house years, City and State of the person. For the dataset, you can use this url Dataset Link
- Income : specifies the monthly salary of the loan applicant.
- Age : specifies the age in years of the loan applicant.
- Experience : overall professional experience of loan applicants in years.
- Marital Status : Whether loan applicant is married or single.
- House-ownership : Does the loan applicant own a house or lives in a rent apartment.
- Car-ownership : Does the loan applicant own a liability like car.
- Current Job Years : specifies the years of experience in current job years.
- Current House Years : specifies the years in current house residence.
- City : This is address feature that describes the city of residence.
- State : It specifies the State/Union Territory of the loan applicant.
- Risk_Flag : It is a binary indicator whether the loan applicant is defaulter or not.
The given repo consists of main.ipynb (jupyter notebook) and Dataset. The complete program is written in main.ipynb
To run the program, you can execute any of the two steps :
-
Run the project program in Local Computer:
- Download the main.ipynb file in your PC.
- Run the main.ipynb cell.
-
Run the project program in Google Colab :
- Download the main.ipynb file in your PC.
- Upload the main.ipynb to Google Colab.
- Run the main.ipynb cell.
Note :
- Do check the requirement.txt file for software versions used in the project.
Task :
- Create requirement.txt