FRAUDULENT RIDE DETECTION USING MACHINE LEARNING AND DATA MINING TECHNIQUES
Loading...
Date
2021
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
UMT, Lahore
Abstract
We see multiple criminal activities in daily life like bank credit fraud detection, fintech, cybercrime, etc. The Ride-Hailing industry is increasing like Uber and Careem with the same ratio of fraudulent activities; for example, fraudulent rides to achieve bonuses are growing. In our circle, many companies are working and providing rides facilities. Several cheaters and fake riders are penetrating day by day. It is challenging for us to check every ride; differentcompanies offer a bonus to drivers/riders to make money. Some drivers make fake/dummy rides to improve metrics like Number of rides, rating, completion rate, login hours, and acceptance rate. Fake rides damage the marketing budget, and customers/riders also disturb. The usage of ride-hailing services like Uber, Careem, bykea, etc., has received significant attention in recent years, increasing the number of fraudsters attempting to exploit these systems. We present a methodology for detecting fraud in ride-booking systems in this research. The fundamental approach to resolve this problem is an example of anomaly identification. Anomaly detection in GPS measures the distance between one point to another point. The longest distance between two points in GPS shows that there is disconnection. In simple words, if the calculated point value is greater than the threshold, it means there is an anomaly otherwise average. This anomaly detection is not a hard and fast rule to define a GPS error or NOT, but it helps us check route anomaly detection. The suggested framework adapts to the fluctuations in data in the ride-booking environment and identifies fraud with high precision. Currently, Uber is using Relational Graph Convolutional Network methodology for fraud detection. Mainly our focus is to detect Fraudulent and fraud-based rides by using ML and Data Mining techniques.