Big Data Collation And Algorithm Mechanism To Set-Up & Maintain Rules Engine
At QPS, data analysts create merchant and customer profiling database (MPD and CPD respectively) through extensive Big Data mining on a weekly/daily/monthly basis. In the next stage, we integrate merchant’s historical transaction data & outcome chargeback (CB) vs. No-chargeback (No-CB) for each transaction. It is conducted with enhanced customer profile & merchant profile and including the fields from external sources & QPS’s IP lookup geo-location tool.
QPS uses advanced decision tree algorithms to mine data driven rules to categorize each transaction in predicted classes of CB and No-CB.
In order to keep the outcome highly optimized and precise, QPS uses Bayesian conditional probability based algorithm to compute CPR (Chargeback Propensity Rate) for each transaction. In the next step, Big Data algorithms outputs are stored in CNP implementation engine.
In the final stage, detailed historical transaction data and outcomes in form of CB vs No-CB are obtained by QPS and on the basis of that classification rules are refined and then mined through our Big Data analytics.
Multi-Layered Big Data Collation From External Sources
- IP Address
External Big Data
- Social Media Data
- IP Address to Geo-location Mapping(City, State, Country, Zip)
- Telephone number verifiers
- Email Information
Big Data: An Effective Contributor At QPS With A Runtime Deployment Environment
- Enhance transaction data using external Big Data on customer profile from Social media and other sources
- MCC and merchant verification using our internal database - MPD and CPD
- Deploy optimized Predictive Classification Rules to classify new transactions and CPR Scoring algorithm to generate CPR Score for them (Rules /Scoring Engine)