The client was a well-established lending firm having a client base across the globe. They were facing challenges in judging their customer’s loan settlement capabilities, further leading to losses and affecting their decision-making abilities.
- The client was not able to judge their customer’s loan repayment capabilities.
- They were encountering challenges in analytics and decision making.
We developed an application to predict loan repayment capabilities & validate the eligibility of potential consumers.
Tools & Technologies Used:
- Languages: Python & React.JS
- EDA: Density plot, Scatter plot, Box plot, Outlier detection on this dataset using Jupyter notebook
- Feature Engineering: PCA, Standardization, Normalization, Dummy feature addition
- Prediction Algorithm: Random forest, Logistic regression, Decision Tree Classifier
- Libraries: Flask, Sklearn, Joblib, Pathlib, Pickel
Impact of the implementation:
- It facilitated the identification process of potential fraud customers, making the process simpler.
- Decrease in the number of potential fraud customers by 70%.
Key Success Factors
Empowering our solutions with AI & ML helped us in creating more sustainable products for our clients.