Research On Data Mining Technology In Improving Alarm Processing Efficiency Of Network Management

The development of the telecommunications network system has brought convenience to our lives,so that "the sea memory confidant,the end of the world" is no longer just a good yearning.With the continuous improvement of the construction of the network system and the formal commercialization of the 5G network system,the alarm information of network management system will surely increase exponentially,which brings great pressure on the operation of network management alarm system.How to realize the compressed storage of the alarm data in the system and mine the value information of the alarm data has become an important research topic.By analyzing the shortcomings of traditional Apriori algorithm,the process of generating frequent itemsets is improved,and the idea of dichotomy is used.The technique of combining and integrating the union realizes the generation of frequent itemsets,which greatly shortens the generation process of frequent itemsets and improves the efficiency of the algorithm.The Hadoop parallel computing framework is used to realize the parallelization of the association algorithm,which further improves the efficiency of the associated data mining.In view of the storage pressure of massive alarm information,this paper proposes a hybrid non-full-quantity encoding compression algorithm,which improves the traditional Huffman compression algorithm by using the characteristics of alarm data with a large number of repetitive attributes.The core idea of the algorithm is to use phrases as The basic unit of data compression combines three modes of Huffman encoding,lexicographic encoding and non-encoding to implement compression on the database.In the process of compression,an attribute is used as a phrase to achieve better compression effect.Because there is only one encoding for a phrase to be compressed globally,you can search the position of the original data in the database directly through the encoding of the search content without decompression,and achieve uncompressed retrieval access to the data.In this paper,the network alarm data is used to verify that the improved compression algorithm has better search performance in addition to lower compression ratio.It also verifies that the correlation algorithm mining efficiency is effectively improved.Finally,the alarm information processing part of the existing network management system is redesigned,the improved compression algorithm and alarm data associations fully utilized to improve the processing efficiency of the alarm data.