Date of Award

5-2012

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Computer Science

First Advisor

Dr. Sanjukta Bhowmick

Second Advisor

Dr. Hesham Ali

Third Advisor

Dr. Parvathi Chundi

Fourth Advisor

Dr. Dhundy Bastola

Abstract

The study of real-world systems, represented as networks, has important application in many disciplines including social sciences [1], bioinformatics [2] and software engineering [3]. These networks are extremely large, and analyzing them is very expensive. Our research work involves developing parallel graph sampling methods for efficient analysis of gene correlation networks. Our sampling algorithms maintain important structural and informational properties of large unstructured networks. We focus on preserving the relative importance, based on combinatorial metrics, rather than the exact measures. We use a special subgraph technique, based on finding triangles called maximal chordal subgraphs, which maintains the highly connected portions of the network while increasing the distance between less connected regions. Our results show that even with significant reduction of the network we can obtain reliable subgraphs which conserve most of the relevant combinatorial and functional properties. Additionally, sampling reveals new functional properties which were previously undiscovered in the original system.

Comments

A Thesis Presented to the Department of Computer Science and the Faculty of the Graduate College University of Nebraska In Partial Fulfillment of the Requirements for the Degree Masters of Science. Copyright 2012 Kanimathi Duraisamy.

Download

COinS