基于模糊关联规则的海量气象数据动态挖掘  

Dynamic mining of massive meteorological data based on fuzzy association rules

作　　者：骆阳 张旗[2] LUO Yang;ZHANG Qi(Zhejiang Meteorological Information Network Center,Hangzhou 310000,China;Zhejiang Meteorological Services Center,Hangzhou 310000,China)

机构地区：[1]浙江省气象信息网络中心,浙江杭州310000 [2]浙江省气象服务中心,浙江杭州310000

出　　处：《电子设计工程》2023年第22期149-152,共4页Electronic Design Engineering

摘　　要：海量气象数据之间存在模糊关联性,且这种模糊关联性难以确定,所以研究基于模糊关联规则的海量气象数据动态挖掘方法。结合EMD和MIC设计时间序列数据去噪算法,对海量气象数据进行去噪处理。搭建基于生成对抗网络与时间指数的GAN-TRTI缺失值补全函数,填补时间序列缺失值。使用模糊关联规则与粒子群优化算法设计海量数据动态挖掘算法,实现海量气象数据的动态挖掘。测试结果表明,在所设计方法的挖掘结果中,样本对于挖掘规则的平均置信度较高,最终稳定在92%左右,平均支持度最终达到90%,说明该方法的挖掘效果好。There is fuzzy correlation between massive meteorological data,and this fuzzy correlation is difficult to determine.Therefore,a dynamic mining method for massive meteorological data based on fuzzy association rules is studied.Design a time series data denoising algorithm combining EMD and MIC to denoise massive meteorological data.Build a GAN-TRTI missing value completion function based on generating adversarial networks and time indices to fill in missing values in the time series.Using fuzzy association rules and particle swarm optimization algorithm to design a massive data dynamic mining algorithm to realize the dynamic mining of massive meteorological data.The test results show that in the mining results of the designed method,the average confidence of the samples in the mining rules is relatively high,ultimately stabilizing at around 92%,and the average support ultimately reaches 90%,indicating that the mining effect of this method is good.

关 键 词：模糊关联规则 气象数据 动态挖掘 粒子群优化算法 缺失值补全 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]