Credit Analysis using Decision Trees and RF

Credit analysis has been a tough job for any of the investing firms around the world. Using data visualizations and forming trends in the analysis thereafter has proven to be a tool with good efficiency. Machine learning models can predict the values and determine the probability of a borrower to default a loan considering his/her previous borrowings and/or other features. We have first cleaned the data and created visualizations to understand correlation between different features. Then, we have applied machine learning models, namely, decision trees and random forest, to make predictions. We have achieved an average precision of 81% with the test dataset when random forest model is applied. The project helped me understand how machine learning models can be applied to our real world problems. Looking forward we plan to refine our features and make some modifications to apply even more complex machine learning models for better precision.

Built With

• machine-learning
• python