A home equity line of credit, or HELOC, is a loan in which the lender agrees to lend a maximum amount within an agreed period (called a term), where the collateral is the borrower's equity in their house (akin to a second mortgage). Because a home often is a consumer's most valuable asset, many homeowners use home equity credit lines only for major items, such as education, home improvements, or medical bills, and choose not to use them for day-to-day expenses. HELOC abuse is often cited as one cause of the subprime mortgage crisis.
This project was to evaluate the risk of Home Equity Line of Credit (HELOC) applications by developing a machine learning predictive model and a decision support system (DSS).
Tool: Python
- data preprocessing including data cleaning, dealing with missing values and categorical values and feature scaling to standardize numeric variables
- machine learning modeling, hyper-parameter tuning and cross validation to evaluate model performance.
Algorithms Used
- Logistic Regression
- Support Vector Machine
- Random Forest
- Gradient Boosting Classifier
- Light Gradient Boosting Machine
- K-Nearest Neighbors
Outcome: lightgbm performed best, achieving 0.7958 auc score and 73% accuracy