Comprehensive Centrality Algorithm for Ranking Important Nodes in Networks

Technology #16507a

Questions about this technology? Ask a Technology Manager

Download Printable PDF

Categories
Researchers
David Tran, M.D., Ph.D.
David Tran, M.D., Ph.D., is a neuro-oncologist and an assistant professor in the Department of Neurosurgery at the University of Florida. He earned his M.D. degree from the Mayo Clinic College of Medicine in 2005. He is a member of the American Medical Association, the American Society of Clinical Oncology, and the Society for Neuro-Oncology. His research interests include novel therapies for brain cancer, mesenchymal factors in gliomagenesis, regulation of the epithelial-mesenchymal transition, tumor dormancy, and cancer genomics.
External Link (neurosurgery.ufl.edu)
Son Bang Le, Ph.D.
Son Bang Le, Ph.D., is a senior research scientist in the Department of Neurosurgery at the University of Florida. He earned his Ph.D. in Lomonosov Moscow State Academy of Fine Chemical Technology, Moscow, Russia. His research interests include system network of tumor development, single cell molecular tracing, and application of computational tools, in particular, deep learning to integrate big data to understand cancer biology.
Managed By
Zahara M. Jaffer
Assistant Director 352-392-8929
Patent Protection

US Patent Pending

Analyzes Networks of All Types by Determining the Importance of Each Node

This network ranking system uses a computational algorithm based on a set of parameters to generate an importance score for each node in a network, thereby establishing a comprehensive centrality measure for use in complex network analysis. Many businesses such as social networks, news distributors, electrical power providers, computer network operators, epidemic researchers, communications services, and others rely on network analysis tools to influence decision making and affect their business procedures. The network analytics market is projected to grow to $2.32 billion by 2020. There are many algorithms for calculating centrality, a node importance measure and one of the most fundamental metrics involved in network analysis. A crucial issue in determining the importance of nodes in networks is the desire to incorporate various parameters known individually to influence network ranking, since existing centrality algorithms do not leverage all available network information. Researchers at the University of Florida have developed nSCORE (network Systems Calculation of Optimal Ranking Engine), a network ranking framework, experimentally validated as highly accurate, that combines many existing parameters into a comprehensive centrality algorithm that produces scores to plainly quantify the importance of nodes in any network. nSCORE can be applied to any field that involves the use of networks because it is designed to take any type of network and node statistics as inputs.

Application

Network analysis engine that determines a hierarchy of nodes according to the importance of each node in the network, maintaining compatibility to networks of all types. Examples include analyses to identify genes responsible for cell type conversions in a gene network, social media users producing the biggest promotional impact in a social network, relevance of news stories within news distribution networks, individuals causing the greatest spread of an infectious disease during an epidemic outbreak within a given population, and countless others

Advantages

  • Generates an importance score for each node in a network, providing a comprehensive centrality algorithm for network analysis
  • Incorporates limitless sets of existing parameters into one scoring framework, fully utilizing other known centrality measures and eliminating the need for prior knowledge about the network
  • Takes any type of network and node statistics as inputs, broadening its use for analysis of any network-based system
  • Ranks nodes in a dynamic network, differing from existing algorithms that use only static networks

Technology

A number of centrality measures reflect the importance of a node in a network, such as degree or “betweenness.” While varied approaches can isolate specific effects on other nodes in the network, nSCORE (network Systems Calculation of Optimal Ranking Engine) automatically synthesizes limitless sets of existing parameters that individually influence network properties and uses an iterative computational algorithm to produce a node importance score drawn from all available node information. Because it takes as inputs data sets describing any network and node statistics, nSCORE is capable of ranking nodes in networks and node statistics of any type, whether known or unknown.