Short Ticketing Detection Framework Analysis Report

Each year, fare evasion costs the UK railway system approximately 240 million pounds [1] with short ticketing, where passengers buy tickets for shorter, cheaper journeys but travel beyond the permitted destination, representing a specific and often undetected aspect of the broader issue. A simple but practical example would be: a passenger travelling from Seaside Station to International Terminus Station via Commuter Hub Station and Financial District Station might purchase two separate tickets (Seaside Station to Commuter Hub Station, and Financial District Station to International Terminus Station) instead of the complete journey ticket, potentially saving money while committing ticket fraud leading to revenue loss for the Train Operating Companies (TOCs). To solve this problem, this comprehensive report provides an in-depth analysis of the short ticketing detection framework developed by researchers Yuyang Miao and Huijun Xing at Imperial College London. This study represents an unsupervised machine learning approach. This work is based on a dataset collected from the UK railway system, including 100 stations' entry and exit data for seven days, with approximately 6.5 million trials of records.