Research Framework for Electronic Fund Transfer(EFT) – Switch Processor Interface (PI)
Developed a homegrown machine learning based research Framework for EFT- Switch Processor interface and card authorization using TAL and TACL, achieving 90% analytic and automated testing support for online transaction processing. Key Feature:
- Devised machine learning models for transaction classification using a bag of word and SVM classifier attaining .89 recall and .91
- precision figures to perform automated analytic based research on the new and old transaction.
- 50 % reduction in the overall testing effort was achieved using the research Framework.
- Automated validation of request & response messages for Acquirer, Issuer and Switch at IPC(Inter-process communication) level.
- The framework also supports different modes of testing Unit, Progression, Regression and online testing with other Network gateways (like Visa MasterCard etc.).
- Inline field checkpoints within IPC message.
- Was awarded Accenture Idea of the year-2014-15 for the project.