Telecommunication sector has considerably large network infrastructure consisting of differentrnnetwork equipment to provide telecom services to the customers. As a service provider,rntelecom operators have to make sure that customers are getting services without interruption.rnTo do so, they built network operation center to monitor the network status with the help ofrna fault management system (FMS). When malfunction or error appears on a piece of networkrnequipment, they typically produce an alarm and transmits to the FMS, which is then shown tornthe monitoring team for trouble ticket creation.rnThere is also a chance that failures that happened on one network element can cause other networkrnelements to produce alarms as long as the network equipmentâ€™s are functionally dependent.rnThus, during the trouble ticketing, the operators may need to collect further informationrnsuch as relationship or dependency between alarms to ultimately distinguish the root source ofrnthe alarms. But the information on FMS lacks explanation and indication of the causes, forcingrnmonitoring teams to manually figure out the cause and take action, which results in a poorrndecision making as well as a longer trouble shooting and service outage time.rnThis study focuses on studying the association between network equipment alarms using a correlationrnanalysis technique called Pearson correlation using history alarm data and studyingrnthe feasibility of using artificial neural network to build fault localization model to support thernmonitoring team with decision making. Fault localization models using one specific neuralrnnetwork architecture are built in different literatures. But in this study two neural network architecturesrnnamely, Feed forward neural network (FFNN) and cascade forward neural networkrn(CFNN), are selected to build a model using the data of correlation analysis output, the microwavernlink topology and the history alarms as an input. Data preparation techniques such asrndata engineering and feature engineering have been applied to the collected alarm data fromrnthe existing fault management system. After building the models, they are evaluated usingrncommonly used performance metrics such as accuracy and error measurement. Python is usedrnas programming tool to perform the correlation analysis as well as to develop and compare therntwo neural network models. Experimentation results exhibit that CFNN achieves an accuracyrnof 97.7% while FFNN achieves 95.9% overall accuracy.