Short Ticketing Detection Framework Analysis Report

TL;DR

This paper analyzes an unsupervised machine learning framework for detecting short ticketing fraud in railway systems, utilizing multiple algorithms and station classification to identify suspicious patterns and fraud behaviors.

Contribution

It introduces a multi-expert unsupervised framework with station classification and multiple algorithms for effective short ticketing fraud detection.

Findings

Identified five distinct short ticketing fraud patterns

Successfully classified stations into high-risk categories

Demonstrated potential for ticketing recovery in transportation systems

Abstract

This report presents a comprehensive analysis of an unsupervised multi-expert machine learning framework for detecting short ticketing fraud in railway systems. The study introduces an A/B/C/D station classification system that successfully identifies suspicious patterns across 30 high-risk stations. The framework employs four complementary algorithms: Isolation Forest, Local Outlier Factor, One-Class SVM, and Mahalanobis Distance. Key findings include the identification of five distinct short ticketing patterns and potential for short ticketing recovery in transportation systems.