2001

Sachinopoulou, Anna

Visualization is a powerful tool for data mining. During the last 30 years, multi-dimensional visualization methods and techniques have been developed in order to conduct data mining (Visual Data Mining) or to efficiently present and evaluate the results of a data mining query. The purpose of this review is to list and shortly describe such methods and give some examples of multidimensional visualization so that the tool designers and data mining analysts can get an overview of the existent possibilities. The methods described are divided into six groups: Geometric techniques where the methods described are based on the Cartesian coordinates; Icon techniques where the main tool and basic unit of the visualization technique is an icon; Pixel Oriented techniques where the pixels are used as the basic visualization unit; Hierarchical techniques where the data items and the query results are presented in hierarchical displays; Distortion techniques where the three-dimensional space is distorted to allow more information to be visualized; and Graph Based techniques where the information is presented using nodes and edges. Examples of visualization tools give a more complete image of the application of these methods.