Feature Aggregation Across Private Datasets
The One Sentence Summary: Financial institutions are able to privately compute across common customers across banking, insurance, and telecom providers to train more accurate models (credit scoring, risk, marketing, etc.) using the combined feature set.
Situation: A large Singaporean bank wants to improve its credit scoring, risk, and marketing models by accessing more data.
Complication: Data sharing initiatives have historically been difficult for the bank to implement for several reasons, including Singaporean data privacy laws and data confidentiality concerns.
Resolution: XOR enables the bank to compute across common customers across banking, insurance, and telecom providers to train more accurate models using the combined feature set.
Outcomes:
-
Internal data scientists have enhanced predictive accuracy of their algorithms
-
Predictive features can be identified and joined across jurisdictions without moving or revealing the underlying sensitive data
CURIOUS TO LEARN MORE ABOUT FEATURE AGGREGATION ACROSS PRIVATE DATASETS?
LET’S TALK!
Please contact our Business Development and Solutions team to learn more about Feature Aggregation Across Private Datasets!