|
Activity Number:
|
31
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Statistical Learning and Data Mining
|
| Abstract - #305780 |
|
Title:
|
Multi-Mode Networks
|
|
Author(s):
|
Walid Sharabati*+ and Yasmin Said and Edward J. Wegman
|
|
Companies:
|
Purdue University and George Mason University and George Mason University
|
|
Address:
|
307 Montifiore St. #300, Lafayette, IN, 47905,
|
|
Keywords:
|
multi-mode networks ; social networks ; infinite vector spaces ; data mining ; multi-dimensional tensor analysis ; relational networks
|
|
Abstract:
|
There has been discussion in the literature about dichotomous (binary) two-mode (bipartite) networks, but no research has yet contributed to the expansion of this problem and its applications. In this paper, we present an advanced mathematical methodology to manipulate multi-mode (higher-dimensional) large-scale weighted networks. In network theory, a network may be represented with a matrix or graph-a set of vertices and edges. We focus our study on the matrix approach; we incorporate multi-dimensional tensor analysis to expand on vertex attributes and infinite vector space analysis to address network expansion in terms of vertices and edges. We then utilize linear algebra to derive relational networks from the infinitely dimensional multi-mode networks. Finally, we show an example of a three-mode network and all its descending lower-mode networks.
|
