Title :
Dimension reduction methods in graph drawing problem
Author :
Zaoralek, Lukas ; Burianek, Tomas ; Snasel, Vaclav
Author_Institution :
Dept. of Comput. Sci. & Dept. of Electr. Eng., VSB - Tech. Univ. of Ostrava, Ostrava, Czech Republic
Abstract :
Graph is a standard data structures mainly used to impose relations in a connected network of objects. Graphs are drawn mostly in ℝ2 space with vertices as points and edges as straight lines connecting vertices. Methods for reduction of dimension are intended for projecting high-dimensional data to a new representation in low-dimensional space. Objective of this work is to use selected dimension reduction methods as a main step in construction of a graph drawing. Graph drawings by selected methods are compared to each other and also to the classical approach presented by Kamada and Kawai.
Keywords :
graph theory; ℝ2 space; dimension reduction methods; graph drawing problem; high-dimensional data projection; low-dimensional space; Computational modeling; Dolphins; Electronic mail; Autoencoder; Dimension Reduction; Graph Drawing; Kamada-Kawai; Neighborhood Preserving Embedding; Sammon´s Mapping; Symmetric Stochastic Neighbor Embedding;
Conference_Titel :
Intelligent Systems Design and Applications (ISDA), 2014 14th International Conference on
Print_ISBN :
978-1-4799-7937-0
DOI :
10.1109/ISDA.2014.7066252