The solution, which has become part of the marketplace's comprehensive anti-fraud system, is based on ML models developed by the Beeline and Wildberries teams, including more than 5,000 parameters — from aggregated information about user actions on the marketplace to data from the Beeline ecosystem.
Machine learning algorithms identify suspicious activity in real time, help block it, and warn the platform about the most complex cases where additional verification is needed.
Wildberries explained:
The use of data analysis at the intersection of telecommunications and e-commerce has made it possible to create an innovative approach to strengthening security measures for the industry – the service reduces moderation costs while increasing the trust of sellers and buyers in the platform.