Machine Learning Techniques for SIM Box Fraud Detection

In today's competitive environment, telecommunication operators and service providers need to generate revenue by designing and delivering innovative services to the subscribers. At the same time, a prime consideration is to minimize the cost and prevent the revenue leakages. In this context, the industry faces numerous challenges and different types of frauds. There is a continuous effort to tackle this problem by improving the implementation methodology and the network protocols. However, in general, fraud detection is difficult and currently addressed in a proactive manner. Fraudulent callers exploit the weaknesses of specific protocol level solutions and avoid detection of the gray traffic. Our research work is inspired by classification algorithms used in machine learning and employed in different fields of science and engineering e.g. images processing, speech recognition, spam email detection etc.. We have applied these machine learning techniques (MLTs) for the classification of normal and fraudulent subscriber (SIM Box). We have used the call detail records (CDRs) of normal and fraudulent subscriber as an input to identify the important attributes; 25 for each customers. These attribute are used for classification of the normal and fraudulent subscribers using Neural Network (NN) and Support Vector Machine (SVM). A comparative performance analysis of both techniques is also presented using various evaluation parameters. SVM using the kernel (Polynomial, Radial, and Sigmoid) show best performance with an accuracy of 99.24%. SVM Linear kernel show the worst performance with accuracy of 95.18 % and 0.19 regression. In case of NN, Bayesian Regularization and Resilient Back-Propagation algorithms show best and worst performance with an accuracy of 99.87 % and 99.53% respectively.