Category Prediction of Financial Fraud Related Posts in Telegram Group Chats
Chunlan Gao, Yubao Wu · 2024
In the last decades, Internet usage has been growing rapidly. However, as the Internet becomes a part of day-to-day activities, criminal activity is also on the rise. Lots of research has been developed to analyze the dark web marketplaces. However, with the development of various chat applications, illicit vendors have gradually shifted from dark web marketplaces to chat platforms. A popular platform is Telegram. The cyber crimes on Telegram encompass various types, with Financial Fraud being one of the significant categories and having the most significant impact on our lives. This paper aims at categorizing different types of financial fraud-related posts on Telegram. This research comprises several works: (i) Downloading chat records related to financial fraud from Telegram channels, and manually categorizing the text data; (ii) Pre-processing of the data and developing models for data analysis and classification; (iii) Comparing the accuracy of different embedding method in different models and conducting an analysis. The main task is to classify these posts by making use of different kinds of feature extraction methods and machine-learning techniques.