Changepoint model for Bayesian online fraud detection in call data

Official URL: http://risc01.sabanciuniv.edu/record=b1817040 (Table of Contents)

Illegal use in the phone network is a massive problem for both telecommunication companies and their users. By gaining criminal access to customers' telephone, fraudsters make an illicit pro t and cause heavy tra c in the call network. After rising trend in mobile phone fraud, telecommunication companies' security departments mainly focused on increasing the e ciency of fraud detection algorithms and decreasing the number of false alarms. In this thesis, we represent an online event-based fraud detection algorithm based on Hidden Markov Models (HMM). Detection problem is formulated as a changepoint model on caller's behavior. To capture call behavior more speci cally, we split it into three parts; call frequency, call duration and call features. We prefer to adapt changepoint model for call data because of its memoryless property; the data before the changepoint does not depend on the data after the change point. To investigate the performance of our algorithm, we conducted an extensive computational study on our generated data. Our results indicate that the algorithm is practical and resampling methods can control the di culty of linearly increasing computational cost.