A Cascading Neural Network Ensemble for Locating Faults of Large-scale Information Systems
Zhifeng Chen, Minjing Peng, Bo Li · International Journal of Security and Its Applications · 2014
Failure types of Information Systems are varied, which leads to the long duration for locating faults, and decreases the quality of customer service or income.In order to accurately and timely locating BOSS faults, a cascading neural network ensemble was proposed.In the proposed ensemble, the approach of ensemble was used to promote the accuracy of locating faults, and the cascading structure was employed to decrease the time duration of locating.The proposed approach used in fault location BOSS achieved the following results: (1) the speed of locating faults has been improved; (2) complaint rate from customers has been reduced; (3) the billing losses have been decreased.