A novel high-dimension data visualization method based on concept color spectrum diagram

摘要

According to the conceptualization habit of human cognition and perceiving characteristics of the human eyesight, a novel high-dimensional data visualization method, which combined uncertain concepts extraction methods and visualization techniques, is proposed in this paper. The technical framework of generating a concept color spectrum diagram of a high-dimension data record is presented. The cloud transformation is applied to extract uncertain concepts, and an approach that represents a concept by a specific color in HSV color space is designed. Hence the color parallel coordinates of cloud sets are established, it could display the distribution features of data while represent corresponding low-dimensional concepts. The concept color spectrum diagrams are generated to provide an overview of data records. All of the visualization processes are illustrated by figures of a test data set. This method has great potential in multivariate data presentation.