Fault diagnosis in mobile computing using TwinSVM

Abstract

Introduction: Mobile computing systems (MCS) comes up with the challenge of low communication bandwidth and energy due to the mobile nature of the network. These features sometimes may come up with the undesirable behaviour of the system that eventually affects the efficiency of the system. Problem: Fault tolerance in MCS will increase the efficiency of the system even in the presence of faults. Objective: The main objective of this work is the development of the Monitoring Framework and Fault Detection and Classification. Methodology: For the Node Monitoring and for the detection and classification of faults in the system a neighbourhood comparison-based technique has been proposed. The proposed framework uses Twin Support Vector Machine (TWSVM) algorithm has been applied to build classifier for fault classification in the mobile network. Results: The proposed system has been compared with the existing techniques and has been evaluated towards calculating the detection accuracy, latency, energy consumption, packet delivery ratio, false classification rate and false positive rate. Conclusion: The proposed framework performs better in terms of all the selected parameters.