Application of Data Science in Fraud Detection
About the speaker
Somenath had pursued a Masters in Computer Science from Banaras Hindu University (BHU). He has 12+ years of experience in Statistical/Machine Learning, Predictive Modelling, Product Management and Analytics Consulting. He conceptualized and delivered actionable models spanning various industries, including Finance, Media, Retail, Wholesale, and Logistics. Currently, he holds the position of Senior Manager - Data Science at AB InBev, where he leads data science initiatives and also owns analytics products.
In this interactive session, Mr. Somenath shares his career journey in Data Science, highlighting significant milestones. He delves into the vital role of Data Science in tackling fraud, with a spotlight on the pressing issue of account takeovers by fraudsters. He provides insights into the architecture considerations and the art of selecting the right models for effective fraud analysis, ensuring accuracy and efficiency in detection. Data exploration's importance is emphasized by him, especially in identifying fraud patterns. Addressing skewed data challenges, he shares strategies to improve model accuracy.
Somenath sheds light on critical data preprocessing techniques, feature engineering methods, the significance of exploratory data analysis (EDA), and robust model validation methods in the fraud analysis process to ensure accuracy. Listeners can gain an understanding of the strategies employed for deploying fraud detection models into production and the ongoing monitoring required to stay ahead of evolving fraud tactics.
Watch the entire video, including the insightful Q&A session at the end to gain industry knowledge.