How We Calculate ROI with Machine Learning
Our ROI calculation leverages a machine learning model trained on last year's territory-specific and loan-related data. Using a Linear Regression model, we analyze correlations among essential features such as loan size, territory performance, borrower risk profiles, loan duration, renewal cycle, customer retention, and loan recurrence. This approach enables us to forecast ROI with precision and relevance.
To align with actual market conditions, certain key metrics are preset in this calculation. These include an average default rate of 6%—reflecting data that ranges between 4.6% and 6%, with 6% used as a conservative measure in this calculator—and a loan turnover rate of 5. These baseline metrics refine the model, enhancing the ROI calculation's realism and alignment with current industry data.
For additional precision, we engineered advanced features such as weighted loan-to-value ratios, segmented borrower profiles based on risk and repayment patterns, and renewal frequency. Hyperparameter tuning and cross-validation techniques were applied to ensure robust and reliable outputs, while continuous model retraining keeps our predictions aligned with changing market trends.
This data-driven, adaptable approach to ROI calculation provides precise, scenario-specific insights tailored to each unique set of circumstances. However, please note that this calculation focuses on core ROI drivers and does not include peripheral costs, allowing for a streamlined, focused analysis.