Network Alarm Analysis Based on Data Mining and LSTM

Abstract

Network alarms are becoming more and more complex, and the scale is increasing sharply. When the root alarm is generated by the devices, a lot of redundant alarms are often triggered. The traditional processing mode is difficult to find the root alarm and filter out the redundant alarm. We provide a new analysis method based on data mining and long short-term memory(LSTM) to find the root alarm and to filter out the redundant alarm.