Method
Graph features
Features aggregated from neighbors and graph structure.
10 cases from 9 companies apply Graph features. 2 were checked in an online A/B test; 5 report results in production.
Typical tasks: Catch fraud and abuse, Moderate content at scale, Decide who gets an offer, Recommend from what the user just did and 4 more. Companies: Yandex, Avito, Airbnb, Alfa-Bank, LinkedIn and 4 more. Often combined with: Fraud scoring, Gradient boosting, Golden set / benchmark, Sessionization, Time-based validation and 14 more.
About Anti-Fraud and Cats: How to Fight Fraud, Not Just Detect It Better![]() | ![]() | general | 2026-05 |
| Solving the Task: Life Hacks from the Task Author (AvitoTech ML Cup 2026) | classifieds | 2026-05 | |
The Social Graph as an Ad Targeting Tool![]() | ![]() | telecom | 2025-05 |
Graphs in Anti-Fraud and How We Apply Them![]() | ![]() | social | 2024-11 |
| Leverage graph technology for real-time Fraud Detection and Prevention | travel | 2024-07 | |
| Not Scoring Alone in the Field: Social Scoring at oneFactorON | ON | fintech | 2024-05 |
| Graph Usage in Combating ATO Fraud Risk | fintech | 2023-09 | |
| Graph Machine Learning at Airbnb | marketplace | 2022-06 | |
| Social Scoring and a Phone-Number-Based Approval Model | banking | 2022-06 | |
| Community building recommender for group chats in LinkedIn Messaging | social | 2022-04 |

