Research And Implementation Of Clustering Analysis Algorithms On Spatial Data

With the increasingly development of the Information Technology and Geographical Information System (GIS), GIS has been widely applied in many different fields. Because of diversification of data capturing technologies and means, people have obtained a great many of data which have relationships with spatial locations. To discover knowledge from huge spatial databases, people urgently need some kinds of powerful analytical tools that can fulfill these tasks. Therefore, under these circumstances, spatial data mining has become more and more popular among researchers.Spatial clustering analysis is one of the main research fields of spatial data mining. Other than discover and extract clustering rules deeply hided in large columns of spatial databases, it can also be used combining with other data mining methods to discover information efficiently and effectively in a deeper level.Firstly, this paper introduces the concepts, functions and processing procedures of data mining and clustering analysis. And then, it simply discusses the spatial data mining and spatial clustering analysis which involving spatial data. Secondly, it introduces some kinds of currently existing clustering analysis methods, not only including the traditional typical spatial clustering analysis algorithms, but also including some new clustering analysis algorithms that have been developed quickly in recent years.Next, this paper discusses hierarchical clustering in detail. In the beginning, an elaboration of concepts and fundamental ideas of hierarchical clustering is made. Then, it introduces the process of clustering analysis using the hierarchical clustering. And then, some typical analysis methods of hierarchical clustering that are commonly used are discussed. At last, the application of hierarchical clustering is analyzed with a real instance in our life.Next, this paper chooses several typical clustering analysis algorithms from numerous clustering analysis algorithms, such as K-means, K-medoids and DBSCAN, and makes some systematical research on them. First, each algorithm is explained in detail, and the illustration of implement process of each algorithm is made. Then each algorithm is implemented via program. Finally, all of these algorithms are analyzed in detail and some conclusions are made, including the performance, complexity, advantages and disadvantages of different algorithms, and so on.At the end of the paper, it gives a summary and an outlook for the research of the clustering algorithms on spatial data, and put forward for improvement areas, while propose research direction and goals to study and research the clustering algorithms on spatial data in future.