基于告警日志分析的移动网络故障预测

In the mobile networks, those massive amount of log data generated by various types of devices is a language which earlier than themachine characteristics thatexpressed by alarms. Make full advantage of log data and take an intensive study of the relationship between the alarm logs and device failures, it will be able to predict network device operation failure in advance. Using big data technology and some Machine learning algorithms to implement log mining analysis and log monitoring warning, so it can gradually turn the traditional monitoring mode to "machine-based monitoring, supplemented by manual intervention" through the discovery of log anomalies,root cause prediction and failure warning, and then achieve the prupose of intelligent, low failure rateand enhance perception. Comparies experiments between Test set and training set show that this method is effective.