Study on visualization of Spatial Data Mining and Knowledge Discovery

摘要

Spatial data usually show as vector maps or raster images. The vector maps are composed of points, lines and polygons, while raster images are composed of grids. Spatial data have different character from common data, which makes it difficult to find knowledge from spatial database than from common database. Spatial Data Mining and Knowledge discovery (SDMKD) is considered as an advanced process to extract novel and interesting models from spatial database. However, visualization is a technique which can deal with substantive data and show this process as figures, images and charts by computers. But, SDMKD and visualization always keep separate and develop independently. A basic and important problem about visualization supporting SDMKD is seldom concerned. To make the process of SDMKD efficient and intuitional, this paper has a preliminary study on visualization of SDMKD through a mutual operation between the human and the computer. It realizes the visualization of data rebuilding and the process of data mining. Some techniques of visualization are used such as charts, maps, figures and images.