Applied Machine Learning for realtime #FairPlay against Fraud
Submitted by Aditya Prasad Narisetty (@adityaprasadn) on Monday, 20 November 2017
For any firm processing online transactions, ensuring a strong shield against fraud is of top priority. And for platforms hosting fantasy sports and online gaming, ensuring a fair play from all users and real-time fraud detection is a first line of defence. Traditionally, rule based engines formed the crux of anomaly and fraud detection. But maintaining a rule engine and adapting to new patterns of abuse are tedious tasks. We’ll see how Machine learning can help us ensure fair play on the platform using user and system finger printing.
- Challenges at Dream11, India’s largest fantasy sports platform
- Referral and promotional events, user registration and game play.
- User data collection and preparing training data
- Regression and Gradient Boosted Models
- Scaling up for real-time decision making
- Business impact and key takeaways
Beginner to Intermediate level of M/L basics
Aditya Prasad Narisetty is a Sr. Data Scientist @Dream11 building data driven products from fraud prevention, User & Revenue estimation, marketing attribution, data pipelines and real-time M/L intelligence. Earlier, he was heading the Data Science team at Craftsvilla building recommendation systems, Data Platform, Search, Autosuggestion, real-time inventory profiling, and Fashion Recognition using CNNs.
He’s an avid speaker in the Mumbai machine learning community presenting at GDG Mumbai‘17, AWS conf‘16, DataNativesX, HYSEA IIT-H, Mumbai AI meetup and a couple of other meetups in Mumbai.