基于StarRocks的Al关联规则挖掘技术研究与应用

打开文本图片集

中图分类号：TP274，TP392 文献标志码：A 文章编码：1672-7274（2025）11-0083-04

Research and Application of AI Association Rule Mining Technology Based on StarRocks

LIWeihong，Wu Shuang1，ZHANG Yong1 ，LONG Qiμ1 ，HUJing，CHEN Meizhi² （1.China Mobile Communications Group Anhui Co.，Ltd.，Hefei 23oooo， China; 2.Bright OceansInter-Telecom BeijingCo.，Ltd.，Beijing10oo43，China）

Abstract： This paper proposes an AI alarm association rule mining scheme based on the new generation of MPP analyticaldatabase StarRocks This scheme utilizes the high-performancedataanalysis capability of StarRocks to efciently process massive historicalalarm data and provide ahigh-qualitydata preparationand feature engineering environment for the associationmining algorithms such as WEclat （frequent itemset miningbasedon vertical data format）and WFP-growth （weighted frequent patern growth）.The study shows thatthe based on StarRocks greatly improves the eficiency and accuracy of alarm association rule mining，achieves the transformation of operation and maintenance mode from "experience-driven"to "data intellgence-"，and provides effective technicalsupport for the intelligent operation and maintenance （AIOps） of communication networks.

Keywords： StarRocks; alarmassociationrulemining;WEcalt algorithm;WFP-growth algorithm

0 引言

通信网络设备在运行过程中会产生大量的告警信息，这些告警并非孤立存在，往往由根因故障引发一系列关联告警。（剩余4771字）