The Research Of Predicting Clustering Analysis Algorithm For Moving Data

Clustering analysis is one of important directions of data mining. The process of clustering data is a process of unsupervised self-study classifying. The research of clustering algorithm is a challenging task. The object of clustering is physical or abstract data, these data can be classified two class: static data and moving data. Clustering analysis for static data started early, it has mature data model and clustering algorithm.. Clustering analysis for moving data is an arising task in 21 A.D. In the process of clustering analysis for moving data, according to the historical data, we can get the different clustering of different period, and then according to these clustering, we can predict the class in the future, so we can predict events(collision or split events). Comparing with static data, because almost all data objects are moving and changing, so moving data have more values of application, for example, it has very important value of application in astronomy, military affairs, transportation and moving communication etc. The clustering analysis for moving data is a very challenging task, it mostly represent: (1) the amount of data is very great, it includes different period data. This is the most difficult aspect of clustering analysis; (2)the structure or model of moving data is a new research aspect, it has no mature theory; (3)the goal of clustering analysis for moving data is to predict the future clustering according to historical clustering, so it is very hard to get accurate result, and it is a difficult task.. But, because clustering analysis for moving data is a new and valuable research domain, it has unmeasurable prospect.The most important point of this article is research of clustering analysis for moving data. Firstly, we analyze, generalize and study important clustering algorithm for static data. Then we study the data model and clustering algorithm in moving environment and provide a new predicting clustering analysis for moving data. This article gets these scores:1. Analyzing popular clustering algorithm for static data from different viewpoints, and providing a new mixed clustering algorithm.We analyze some main clustering algorithm for static data from clustering criteria, cluster representation and algorithm framework etc. According to these, we provide a new mixed clustering algorithm- MST Clustering Algorithm Based on Grid. This algorithm optimize MST Clustering Algorithm by grid data, it can efficiently identify...