Working Paper — 2025

BASIC: Behavioural Analytics for Score Improvement in Credit

A Multi-Modal Machine Learning Framework for Credit Score Prediction, Bureau Sensitivity Estimation, and Personalised Improvement Pathways in India

Shubham Arawkar·Pranav Murali·Anupam Acharya·TARA AI Labs
550,000
Users in Training Set
10M+
Sample-Months
92%
Prediction Accuracy
94%
Directional Accuracy

Paper Highlights

  • First population-scale credit score improvement framework for India
  • Integrates bureau tradeline data, banking cashflow signals, and behavioural telemetry
  • Causal impact estimation using Causal Forests and Double Machine Learning
  • Bureau Sensitivity Module recovers latent scoring sensitivities from observed data
  • Median score improvement of +52 points over 90 days for recommendation-followers
  • Trained on user-consented data under DPDP-compliant protocols

Request Access to Full Paper

This paper is shared on a case-by-case basis with researchers, industry professionals, and institutions working in credit intelligence, financial inclusion, and responsible AI. Fill out the form below and our team will review your request.

Your information will only be used to process this access request. We don't share your data with third parties.