The growth of electronic toll collection has led to more fraud, especially when serious drivers use FASTags to ease traffic congestion to avoid repeat fraud. This study highlights the urgent need for effective enforcement of FASTag fraud. Our system is designed to detect and prevent false phone numbers by analysing key features such as money transfer, vehicle type, phone line type, and location. Construct complex fraud models using advanced techniques in machine learning algorithms: KNN, XG Boost, Logistic Regression, Decision Tree, Random Forest, Naive Bayes Classifier, Gaussian Naive Bayes, ANN, SVM Classifier etc. This model has shown great success with ANN achieving the highest accuracy (99.88%) and F1 score (99.93%). It helps in creating a safer and more reliable system by increasing the validity of FASTag transactions. By using these algorithms, the aim is to reduce fraud and increase the integrity and security of the FASTag system. In addition to increasing the security of electronic toll collection, the system also offers solutions that will reduce FASTag fraud and ensure fairness in payments across vehicle models.
Advanced Machine Learning Approaches for Fastag Fraud Detection
04.03.2025
458442 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Effectiveness of FASTag System for Toll Payment in India
IEEE | 2021
|Development of Toll Equivalency Factors for FASTag Lanes Under Mixed Traffic Conditions
Transportation Research Record | 2022
|FRAUD DETECTION DEVICE, FRAUD DETECTION SYSTEM, FRAUD DETECTION METHOD, AND PROGRAM
Europäisches Patentamt | 2023
|FRAUD DETECTION DEVICE, FRAUD DETECTION SYSTEM, FRAUD DETECTION METHOD, AND PROGRAM
Europäisches Patentamt | 2024
|