摘要
流数据是近年来关注比较多的一种数据形式,但由于它自身的特点,无法使用传统的算法对它进行聚类分析。数据挖掘是从大规模数据库中提取感兴趣的信息。聚类是数据挖掘的重要工具,它根据数据间的相似性将数据库分成多个类,每类中数据要求尽可能相似。针对流数据的特点,引入一种采用渔夫捕鱼策略的新的聚类算法。该算法采用动态多点随机投鱼网方法,并且根据捕鱼环境的不同采用不同的探测策略。流数据聚类的捕鱼算法是一种即时更新模型的在线聚类算法。
Streaming data, as an important data model, has got broad attention in recent years. But due to its u- niqueness, traditional clustering algorithm cannot be applied on it. Data mining is used to draw interesting information from very large databases. Clustering plays an outstanding role in data mining applications. Clustering is a division of da- tabases into groups of similar objects based on the similarity. A novel clustering algorithm of using the strategy of fisher＂ fishing is introdueed based on the characteristic of streaming data. A dynamie multipoint and random fishing algorithm is used, and different detection strategies are chosen according to different fishing circumstances. Fishing algorithm of streaming data clustering is an online clustering algorithm with timely updated model.
出处
《成都师范学院学报》
2014年第1期122-124,共3页
Journal of Chengdu Normal University
关键词
流数据
数据挖掘
聚类
捕鱼算法
streaming data
data mining
clustering
fishing algorithm
