The Study Of Clustering Algorithm In Case Data Mining

With the development of science and technology, we have accumulated large amounts of data in various industries. However, when faced with such a huge data, we often feel confused. We can not extract useful knowledge for us from these data, resulting in "Data rich and Information poor". In this case, the data mining technology came into being, it can help us extract the valuable knowledge model for us from the large amounts of data. The data mining technology is considered one of the most promising key technologies.Cluster analysis is an important feature of data mining. In recent years, cluster analysis develop rapidly, there has been a lot of cluster analysis, for example, Partition-based approach, Hierarchical clustering approach, Density cluster approach, Model-methods etc. These methods have good effect when they deal with genera issues. However, the traditional cluster analysis techniques are mainly for structured and unstructured types of data. There are little studies based on the cases of semi-structured. Therefore, this paper proposed a new clustering analysis of cases based on the traditional clustering techniques.First, this article briefly introduces the concept of data mining, functions, methods and techniques, and the data mining application in real life. Then, we introduce the basic knowledge of the case representation and case base structure, including case representation based on relational databases and case representation based on XML. Third, we introduce the concepts of cluster analysis, including the definition of clustering, steps, the data structure and the data types, clustering methods and the typical requirements of clustering. Finally, we proposed smooth stitching clustering algorithm based on geometric smoothness based on the expression of the previous section and the basic knowledge of cluster analysis and described the method in detail.