Multiple points representation model of multivariate data
Yonghong Xu - · Journal of Yanshan University · 2008
The multiple points representation model is proposed in this paper. The attribute variables are represented by multiple plane vectors. Every plane vector is formulated by a complex function, and the dataset is expressed as a complex function matrix. Multiple graphical methods such as parallel coordinates, star glyph, constellation graph can be unified by this model. The repres- entation models' parameters can be optimized by machine learning algorithms so as to reveal the internal structure of original data and serve as a tool for visual pattern recognition.