Application for frequent pattern recognition in telecommunication alarm logs [articol]

Abstract

Based on an algorithm for frequent pattern recognition, this paper presents the implementation of a software application and its respective results in analyzing real-time telecommunication alarm logs. The software application was developed in OMNeT++ (Objective Modular Network Testbed in C++) simulation environment using ACE (Adaptive Communication Environment) toolkit. Different working scenarios are presented in order to simulate extensions of the frequent pattern recognition algorithm: the introduction of time-constraints between alarms and the construction of a Petri net whose transitions are labeled by recognized frequent patterns of alarms.