Skip to content

elijah-medina/car-loan-optimization

Repository files navigation

An alternative to Nomis solutions on e-Cars case through Logistic Regression with Lasso regularization.

Machine Learning mini-project by Ria Ysabelle Flora, James Bryan Labergas, Maria Angela Legaspi, & Elijah Justin Medina

Logistic regression (with lasso regularization) for classification was chosen to predict the Outcome of loan availment since it resulted to the best model accuracy across all car types, namely, New (92%), Used (78%) and Refinanced (68%). The consistent top predictors for New and Used cars are Amount, Rate and Term while Previous Rate went important for Refinanced cars. Among these predictors, only Rate is controllable for e-Car. Using this model, the revenue and captured market is optimized by adjusting the loan rates.

Report and analysis

If you have any questions regarding this study or wish to have a copy of the report, please send me a message via e-mail or LinkedIn. The code used for this analysis is available here with supplementary code here.

About

An alternative to Nomis solutions on e-Cars case through Logistic Regression with Lasso regularization

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published