## A Framework for Physically-Based Information Visualization (1997)

Citations: | 9 - 3 self |

### BibTeX

@MISC{Sprenger97aframework,

author = {T. C. Sprenger and et al.},

title = {A Framework for Physically-Based Information Visualization},

year = {1997}

}

### OpenURL

### Abstract

The following paper describes a framework for the visualization and analysis of economic data. It can be employed in the context of risk analysis, stock prediction and other tasks being important in the context of banking. The system bases on a quantification of the similarity of related objects, which governs the parameters of a mass-spring system, organized as two concentric spheres. More specifically, we initialize all information units onto the surface of the inner sphere and attach them with springs to the outer sphere. Since the spring stiffnesses correspond to the computed similarity measures, the system converges into an energy minimum, which reveals multidimensional relations and adjacencies in terms of spatial neighborhoods. In order to simplify complex setups we propose an additional clustering algorithm for postprocessing. Furthermore, depending on the application scenario we support different topologic arrangements of related objects. In addition, we implemented various interaction techniques allowing semantic analysis of the underlying data sets. The versatility of our approach is illustrated by two examples, namely a comparison of agricultural productivity and an analysis of the relation between interest rates and other economic data.

### Citations

2928 |
Introduction to Statistical Pattern Recognition, Electrical Science Series
- Fukunaga
- 1972
(Show Context)
Citation Context ...around groups of objects. The transparency of the hyperspheres was controlled as a function of the distance to the viewer. Unlike this approach we propose a K-means and PCA based clustering mechanism =-=[19]-=- which will be explained in the upcoming section. The basic idea is to wrap ellipsoids around each cluster whose shape is controlled by the principal components of the respective cluster. The method i... |

112 | A fast adaptive layout algorithm for undirected graphs
- FRICK, LUDWIG, et al.
- 1995
(Show Context)
Citation Context ...s upon relaxation. These types of multidimensional visualization methods have been studied extensively in graph theory, and efficient algorithms had been introduced for fast graph relaxation, such as =-=[11]-=- and [12]. The work reported in our paper was inspired by the research summarized above. However, unlike existing methods, our approach was mostly application driven, where the context was focussed on... |

71 |
Leiserson C, Rivest R. Introduction to Algorithms
- Cormen
- 1990
(Show Context)
Citation Context ...Boolean matrix, whose elements are either 1, if two nodes are connected or 0 if not. Since these types of algorithms belong to the standard repertoire of computer science textbooks, we refer to those =-=[10]-=-. 4.2 Clustering and Level-of-detail (LOD) In order to simplify the geometry and topology of complex object arrangements it is necessary to provide an efficient level-of-detail strategy. Initial work ... |

33 |
Interactive graph layout
- Henry, Hudson
- 1991
(Show Context)
Citation Context ... the following set of equations. We start from the implicit equation of the unit sphere with surface vector x j : | (16) and perform a subsequent affine mapping by ; : surface vector of the ellipsoid =-=(17)-=- Due to the statistical properties of the principal components it is not guaranteed that all K j Ksm j 1 n j ----- r i i 1 = ns= M j 1 n j ----- r i m j -- ( ) r i m j -- ( ) T i 1 = n js= M j u k s k... |

27 |
Research report: Interacting with huge hierarchies: beyond cone trees
- Carriere, Kazman
- 1995
(Show Context)
Citation Context ...various interesting approaches can be found in literature and excellent surveys are available [20]. [5], [14] for instance, visualized text documents and clusters as galaxies and themescapes, whereas =-=[4]-=- proposed cone trees which specifically address hierarchical organization. Another promising method is [12] or [16], who essentially used selforganizing schemes and neural networks to arrange informat... |

26 |
Fast interactive 3-D graph visualization
- Bruss, Frick
- 1995
(Show Context)
Citation Context ...14] for instance, visualized text documents and clusters as galaxies and themescapes, whereas [4] proposed cone trees which specifically address hierarchical organization. Another promising method is =-=[12]-=- or [16], who essentially used selforganizing schemes and neural networks to arrange information objects of the WWW. In a more general understanding, multidimensional visualization problems have been ... |

25 | Visualizing Information on a Sphere - Gross, Sprenger, et al. - 1997 |

17 |
et al., â€œVisualizing the Nonvisual: Spatial Analysis and Interaction with Information from Text Documents
- Wise
- 1995
(Show Context)
Citation Context ...etail control are needed. Due to the importance of information visualization for many applications, various interesting approaches can be found in literature and excellent surveys are available [20]. =-=[5]-=-, [14] for instance, visualized text documents and clusters as galaxies and themescapes, whereas [4] proposed cone trees which specifically address hierarchical organization. Another promising method ... |

16 |
Animating Multidimensional Scaling to Visualize NDimensional Data Sets
- Bentley
- 1996
(Show Context)
Citation Context ...hematical projection algorithms were introduced to map data into subspaces, while preserving their most important features. Interestingly, many current methods use physically based paradigms, such as =-=[18]-=- or [13], where information units are taken as nodes of some generalized mass spring system revealing the structure of relations upon relaxation. These types of multidimensional visualization methods ... |

14 | Visualization of multidimensional shape and texture features in laser range data using complex-valued Gabor wavelets
- Gross, Koch
- 1995
(Show Context)
Citation Context ...tially used selforganizing schemes and neural networks to arrange information objects of the WWW. In a more general understanding, multidimensional visualization problems have been stressed in [6] or =-=[1]-=-. Here, mathematical projection algorithms were introduced to map data into subspaces, while preserving their most important features. Interestingly, many current methods use physically based paradigm... |

11 |
Algorithms & Data Structures: with Applications to Graphics and Geometry
- Nievergelt, Hinrichs
- 1993
(Show Context)
Citation Context ...vector (3) Note, that C also reflects whether or not two objects are related. Hence, it can be considered as some sort of generalization of the so-called adjacency matrix, wellknown from graph theory =-=[8]-=-. The similarity defined by (1) can be interpreted as the projection of one score vector onto another. Needless to say that the quality of the selected key words and retrieval algorithms is crucial fo... |

6 | F.R.CARLS et al., Simulating facial surgery using finite element models - KOCH, GROSS - 1996 |

5 |
et al.: HyperSpace: Web Browsing with Visualisation
- Wood
- 1995
(Show Context)
Citation Context ...instance, visualized text documents and clusters as galaxies and themescapes, whereas [4] proposed cone trees which specifically address hierarchical organization. Another promising method is [12] or =-=[16]-=-, who essentially used selforganizing schemes and neural networks to arrange information objects of the WWW. In a more general understanding, multidimensional visualization problems have been stressed... |

4 |
et al.: Case Study - Narcissus: Visualizing Information
- Hendley
(Show Context)
Citation Context ... order to simplify the geometry and topology of complex object arrangements it is necessary to provide an efficient level-of-detail strategy. Initial work for information visualization is reported in =-=[7]-=- who accomplished simple clustering by wrapping hyperspheres around groups of objects. The transparency of the hyperspheres was controlled as a function of the distance to the viewer. Unlike this appr... |

3 |
A Linear Time Layout Algorithm for Visualizing High-Dimensional Data
- Chalmers
- 1996
(Show Context)
Citation Context ... control are needed. Due to the importance of information visualization for many applications, various interesting approaches can be found in literature and excellent surveys are available [20]. [5], =-=[14]-=- for instance, visualized text documents and clusters as galaxies and themescapes, whereas [4] proposed cone trees which specifically address hierarchical organization. Another promising method is [12... |

2 |
Glassner Principles of Digital Image Synthesis
- S
- 1995
(Show Context)
Citation Context ...e have to be found. A straightforward positioning could be at random. However, due to the problems arising with small numbers of objects, we recommend a Poisson disc sampling procedure of the surface =-=[9]-=-. As depicted in Figure 2b each object is assigned to a solid Fig 2. a) Initialization of information objects on a virtual sphere b) Poisson disc sampling for initial positioning c) Poisson disc distr... |

2 |
Drew: Visualisation of complex systems, http:// www.cs.bham.ac.uk/~rjh/aig.html
- Hendley, S
- 1995
(Show Context)
Citation Context ...l projection algorithms were introduced to map data into subspaces, while preserving their most important features. Interestingly, many current methods use physically based paradigms, such as [18] or =-=[13]-=-, where information units are taken as nodes of some generalized mass spring system revealing the structure of relations upon relaxation. These types of multidimensional visualization methods have bee... |

1 |
Visualizing Strucutre in High-Dimensional Multivariate Data
- Young, Rheingans
- 1991
(Show Context)
Citation Context ...o essentially used selforganizing schemes and neural networks to arrange information objects of the WWW. In a more general understanding, multidimensional visualization problems have been stressed in =-=[6]-=- or [1]. Here, mathematical projection algorithms were introduced to map data into subspaces, while preserving their most important features. Interestingly, many current methods use physically based p... |

1 | Visualisation of complex systems, http://ww.cs.bham.ac.uk - Hendley, Drew |