Research And Design Of Alarm Association Rules Based On Data Mining

In recent years,the scale of the communication network system has become larger and larger,and the structure has become more and more complex,which has led to more and more difficult management of communication network faults and higher management costs.At the same time,data mining technology has gradually entered the public's field of vision,and complex problems have begun to be solved by humans using data mining technology.In view of many considerations,this article plans to use data mining technology to solve the network failure problem.The main research contents are as follows:The initial data in this article is the historical alarm log in the server of the Information Department of Guizhou Provincial People's Hospital.The original data is sampled and selected part of the alarm log,statistical processing,the original alarm log is classified by the classification process application warning,application error,system warning,system error classification,and sequentially imported into the Excel table.The data is pre-processed by EXCEL,and the repeated alarm data and missing data are manually deleted in a short time.The frequency of each type of alarm data and the specific information of the alarm are counted.At the same time,the new alarm data is sorted in chronological order,the relationship of the alarm log in chronological order is observed,and the unit time,alarm type,and alarm event ID are selected as the parameters for association rule mining.The language used for data preprocessing is Python.Time is standardized,and alarm data is sorted in chronological order.The required attributes of the correlation rule are selected,the alarm type and alarm event ID are retained,and the irrelevant attributes of the correlation rule are deleted.The unit time threshold is set,and the alarm-related transaction database is obtained.The association rule algorithm Apriroi algorithm is selected in this paper for mining experiments.Some alarm examples were selected to introduce the principles of Apriori algorithm mining.Parameters such as unit time,support threshold,confidence threshold,and promotion were selected for experimentation and detailed adjustment of the parameters.The association rules and strong association rules under a reasonable threshold are searched,and the corresponding association results are output and analyzed.Associated pre-alarms are used as early warnings for post-alarms.Reasonable alarms are merged according to the result of strong association rules.Post-alarms are deleted and pre-alarms are retained.This is conducive to optimizing the data,and is also conducive to efficiently locating the cause of the alarm failure.Finally,the alarm association rule system is designed for the research content of this paper,and the effectiveness of the system is verified.The content of the system design includes demand analysis and interface design,and the results are displayed on the front-end interface.