Bank-Defaulter
The aim of this data science based project is to find the probability of a person defaulting out of their payment (loans and/or interests).
Input Trainig Data
We are provided with an excel sheet containing the full data of each person , which includes information like :
- Loan ID
- Due amount
- Due Mortgage
- Value
- Reason
- Time
- Past History
- Criminal History
- Defaulter or not
Training Models and Algorithms
Like all other data science problems,my main approach to solve the problem was to start training the data for various types of algorithms, which inlcude:
- Linear Regression
- Logistic Regression
- SVM(Support Vector Machines)
- Ensemble Training
- MLP (Multi-Layered Perceptron)
- Random Forest
And then I checked the accuracy and ROC for each of the algorithsm and compared them.