FINTECH COMPANY
BUSINESS PROBLEM
This FinTech Company developed and optimized a gradient boost algorithm using SAS Enterprise Miner to assess the creditworthiness of their customers. However, they were unable to implement the SAS program in their production environment. To address this challenge, the company needed to convert the SAS code into Python for seamless integration into their enterprise production system.
SCALESOLOGY IN ACTION
The FinTech Company provided Scalesology with an anonymized test dataset that included the input data for their model and the corresponding predictions. They also shared the optimized XGBoost model scoring code generated by SAS Enterprise Miner, which comprised over 11,000 lines of SAS code for a complex gradient-boosting decision tree.
Scalesology successfully converted this program into Python, utilizing Pandas data frame structures and if/then logic. Specifically, the team rewrote:
Data Ingestion and Transformation: The process for parsing input data from preliminary credit attributes into bins.
Scoring Algorithm: The implementation of hundreds of decision tree leaf nodes for the model’s scoring mechanism.
RESULT
In a short time, the FinTech Company received Python code ready for deployment in their production environment. As a result, their gradient boost algorithm could be seamlessly integrated into their enterprise application, enabling real-time credit risk evaluation for their customers.
SERVICE REFERENCES
Comments