The Client
Global Financial Institution
The Challenge
This financial institution required high-performance model management and execution platform in order to run key Loss Forecasting models in preparation for their CCAR (Comprehensive Capital Analysis & Review) submission to the Fed. The client was unable to execute several of their advanced Forecasting models to performance issues. At times the models would take weeks to execute and at times remain incomplete.
The Solution
- Optimized a Markov Chain Monte Carlo simulation based auto loan model
- Addressed performance problems and demonstrated improvements of over 500x on model execution compared to traditional statistical tools implementation
- Instituted model management and repository for better regulatory controls
The Approach
- Implemented Decision Analyzer on HANA; a model management and execution platform that allows the execution of advanced predictive analytics models against multiple economic scenarios
- Realized a high-performance Markov Chain Monte Carlo simulation based auto loan model utilizing Decision Analyzer platform on HANA
- Instigated a central model management environment to manage CCAR models
- Executed advanced visualization and reporting environment utilizing HANA and Explorer