Alarm monitoring method and system, computer readable storage medium and electronic equipment

Abstract

The embodiment of the invention provides a neural network alarm monitoring method and system based on time sliding window alarm verification constraint, a computer readable storage medium and electronic equipment. Data preprocessing is performed by using historical system and equipment alarm data, data cleaning and data conversion of dirty data are realized, system and equipment alarm training sets are made, the system and equipment alarm data sets are trained by using a neural network objective function, a loss function value is optimized, and a system and equipment alarm prediction model is established. Current state information of a system and equipment is collected, the current state information is input into an alarm prediction model to realize alarm prediction, association and aggregation on current and predicted alarms are carried out by utilizing an FP-Growth algorithm, time sliding window alarm verification is carried out according to current network alarms, if verification alarms are wrong, neural network continuous optimization is carried out to adjust to obtain an optimal weight and an optimal offset value, and if the verification alarm is correct, alarm classification, graded output and database backup are carried out.

Description

【Technical field】 [0001] The invention relates to the technical field of communication network fault alarm monitoring and prediction, in particular to a neural network alarm monitoring method, system, computer-readable storage medium and electronic equipment based on a time sliding window alarm verification constraint. 【Background technique】 [0002] With the rapid development of communication and information technology, telecommunication networks are becoming more and more abundant, but the scale and complexity of the network are also becoming more severe. Carrier-class networks usually include thousands of systems and devices from multiple vendors. Systems and devices are interconnected through multiple media, and there are complex relationships among them. This makes fault alarm management a difficult problem in telecom networks. [0003] In the existing operator monitoring solution, the operator has not yet carried out alarm prediction and monitoring through real-time da...

AI Technical Summary

Problems solved by technology

[0008] Network fault prediction based on alarm logs: This technical solution requires the use of historical data and log data, and is highly dependent on historical data. If the historical and log data are lost or incomplete, the accuracy of the technical solution’s prediction results will be greatly reduced; Statistical characteristics of alarm events in the time window to establish a classification prediction model to realize alarm prediction, strong coupling with the state of the system and equipment running time window, relatively poor independence

[0009] Sequential pattern mining to realize network alarm prediction: This technical solution is to use the entire network topology for data feature statistics or data mining. As a result, the accuracy of the prediction results of this technical solution is greatly reduced. This technical solution is too strongly coupled with the network topology, and its flexibility and applicability are low. It is not suitable for complex and flexible networking in the existing network.

Images