基于FP-Growth算法的地面气象观测数据异常挖掘  

Anomaly mining of ground meteorological observation data based on FP-Growth algorithm

作　　者：许烨 牛淑丽 狄增文 Xu Ye;Niu Shuli;Di Zengwen(Zhangye Meteorological Bureau,Zhangye 734000)

机构地区：[1]张掖市气象局,张掖734000

出　　处：《气象水文海洋仪器》2025年第1期33-36,共4页Meteorological,Hydrological and Marine Instruments

基　　金：甘肃省自然科学基金(21JR7RA711);甘肃省气象局气象科研项目(Ms202110,2122rczx十人计划)资助。

摘　　要：为了提高对地面气象观测集合中异常数据的精准检测与识别能力,文章提出基于FP-Growth算法的地面气象观测数据异常挖掘方法。设定数据采集频率,根据观测需求确定数据采集的时间间隔,进行采样地面气象观测数据的整合;引进FP-Growth算法,基于FP-Tree结构,筛选频繁项,进行观测数据特征的提取;对于数据集中的每个点,计算与其最近邻的距离,根据距离定义异常分数,实现异常数据挖掘与聚类。实验结果表明:设计方法挖掘的观测数据异常量与实际样本数据的数量一致,说明该方法在实际应用中,可以实现对地面气象观测数据异常的精准挖掘。In order to improve the accurate detection and recognition ability of abnormal data in ground meteorological observation sets,a ground meteorological observation data anomaly mining method based on FP-Growth algorithm is proposed.Set the frequency of data collection,determine the time interval for data collection based on observation needs,and integrate ground meteorological observation data for sampling;Introducing the FP-Growth algorithm,based on the FP-Tree structure,to screen frequent terms and extract features from observed data.For each point in the dataset,calculate the distance to its nearest neighbor,define the anomaly score based on the distance,and achieve anomaly data mining and clustering.The experimental results show that the abnormal amount of observation data mined by the design method is consistent with the actual sample data,indicating that this method can achieve accurate mining of ground meteorological observation data anomalies in practical applications.

关 键 词：FP-GROWTH算法 挖掘方法 异常 观测数据 气象 地面 

分 类 号：TM933[电气工程—电力电子与电力传动]