Study On Data Visualization Methods And Related Techniques For Clustering

Data mining is a data processing approach that extracts hidden, unknown, potentially useful knowledge and pattern from huge volume of data. It is a new technique that appeared in 1990s for solving the problem of "rich data but poor information". But the procedures of data mining often spent many times and the algorithms of data mining are relatively complex, which result in giving unuseful knowledge and easily coming forth deviation and error. Therefore, the users need to explore some effective methods to more clearly observe the distributing structure of data, to know correlation between data and their developing trends, to understand the procedures and results of data mining. Data visualization technique is a novel, effective approach to resolve that problem and this technique has become the hot spot in the research field of data mining. Data visualization technique is to utilize graphic or image, such as scatterplot, treemap, curve, surface, to display multidimensional non-spatial data, to utilize directly viewing image to help users to explore data set, to deepen the understand of data for users, to quicken the speed for obtaining knowledge.Cluster analysis is an important function in the data mining techniques; especially it has superiority for multidimensional data. As clustering algorithm, the disseration summarizes and analyzes some existing data visualization techniques, and in detail studies the novel data visualization techniques based on clustering, the technique that visualize the procedures and results of clustering analysis and the interactive techniques in the visual clustering analysis etc. This disseration made major works as follows:(1) Although many ideas are suggested in terms of date visualization, there is still no clear definition of date visualization. This disseration made a clear distinction between visualization, Data Visualization, scientific computing visualization, information visualization and their application areas. This disseration systematically introduces the main method of multidimensional data visualization. Their common goal is to display the data attributes and their relationships as much as possible on the limited screen to reflect the relationships among the data. Through this we can have a better understanding of the function in multidimensional data visualization, its utilizing area and its unique...