Leveraging Machine Learning and Image Analysis to Prevent Off - Platform Transactions in Online Marketplaces

Vinay Kumar Yaragani

Abstract: In online marketplaces, preventing off - platform transactions is crucial to maintain transaction integrity and ensure platform compliance. Many users attempt to bypass marketplace rules by sharing contact information to avoid transaction fees or to commit fraudulent activities. Although marketplaces have measures to detect such activities in text - based communication, some users creatively embed contact information directly into product or listing images. This paper presents a comprehensive approach to detect and prevent such listings using a two - step process. First, Optical Character Recognition (OCR) is employed to extract text from the images, converting it into machine - readable content. The second step involves a verification mechanism that analyzes the extracted text for patterns indicative of contact information, such as phone numbers, email addresses, or social media handles. By combining OCR technology with rule - based and machine learning techniques, the proposed approach effectively identifies and takes down listings that violate marketplace policies. This methodology aims to enhance marketplace integrity, reduce off - platform fraud, and safeguard both buyers and sellers.

Keywords: Off - Platform Transactions, Image - Based Contact Detection, Optical Character Recognition (OCR), Marketplace Fraud Prevention, Machine Learning