Get in touch with us at

We would love to hear from you

Anomaly detection model to predict frauds and defaults for a leading payments bank

About Client

Client is a payment and fintech company based in Mumbai. The bank was incorporated on 4th April, 2017 and is a leading phygital bank with 300+ branches and around 2 Lakhs banking points across the country. Its products include savings bank accounts, loans, recurring deposits, remittances, insurance and government disbursements. The company is also an active micro lender in India and currently has a customer base of more than 20 million.


Operationalization and Impact

  • Operationalization of the model in the Big Data Architecture was performed using Pyspark and Shell scripting where model was deployed in the Client’s production environment
  • A scheduler was created in order to give “anomaly score” at regular intervals.
  • Model accuracy: Top 1.11% scorers capture 43% frauds and Top 10% scorers capture 81% frauds.
  • A list of accounts with high anomaly score in every prediction cycle was shared with Client’s Anti-Fraud Unit for verification
  • Model performance tracker was also created and retraining was automated, in the case where accuracy dropped down a certain threshold
Want to learn more about TransOrg’s value proposition, solution methodology and implementation approach?