skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Trend Motif: A Graph Mining Approach for Analysis of Dynamic Complex Networks

Abstract

Complex networks have been used successfully in scientific disciplines ranging from sociology to microbiology to describe systems of interacting units. Until recently, studies of complex networks have mainly focused on their network topology. However, in many real world applications, the edges and vertices have associated attributes that are frequently represented as vertex or edge weights. Furthermore, these weights are often not static, instead changing with time and forming a time series. Hence, to fully understand the dynamics of the complex network, we have to consider both network topology and related time series data. In this work, we propose a motif mining approach to identify trend motifs for such purposes. Simply stated, a trend motif describes a recurring subgraph where each of its vertices or edges displays similar dynamics over a userdefined period. Given this, each trend motif occurrence can help reveal significant events in a complex system; frequent trend motifs may aid in uncovering dynamic rules of change for the system, and the distribution of trend motifs may characterize the global dynamics of the system. Here, we have developed efficient mining algorithms to extract trend motifs. Our experimental validation using three disparate empirical datasets, ranging from the stock market, worldmore » trade, to a protein interaction network, has demonstrated the efficiency and effectiveness of our approach.« less

Authors:
; ;
Publication Date:
Research Org.:
Lawrence Livermore National Lab. (LLNL), Livermore, CA (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
957591
Report Number(s):
UCRL-PROC-231357
TRN: US201016%%41
DOE Contract Number:
W-7405-ENG-48
Resource Type:
Conference
Resource Relation:
Conference: Presented at: The 7th IEEE International Conference on Data Mining (ICDM'07), Omaha, NE, United States, Oct 28 - Oct 31, 2007
Country of Publication:
United States
Language:
English
Subject:
99 GENERAL AND MISCELLANEOUS; 59 BASIC BIOLOGICAL SCIENCES; ALGORITHMS; DISTRIBUTION; EFFICIENCY; MARKET; PROTEINS; SOCIOLOGY; TOPOLOGY; VALIDATION

Citation Formats

Jin, R, McCallen, S, and Almaas, E. Trend Motif: A Graph Mining Approach for Analysis of Dynamic Complex Networks. United States: N. p., 2007. Web.
Jin, R, McCallen, S, & Almaas, E. Trend Motif: A Graph Mining Approach for Analysis of Dynamic Complex Networks. United States.
Jin, R, McCallen, S, and Almaas, E. Mon . "Trend Motif: A Graph Mining Approach for Analysis of Dynamic Complex Networks". United States. doi:. https://www.osti.gov/servlets/purl/957591.
@article{osti_957591,
title = {Trend Motif: A Graph Mining Approach for Analysis of Dynamic Complex Networks},
author = {Jin, R and McCallen, S and Almaas, E},
abstractNote = {Complex networks have been used successfully in scientific disciplines ranging from sociology to microbiology to describe systems of interacting units. Until recently, studies of complex networks have mainly focused on their network topology. However, in many real world applications, the edges and vertices have associated attributes that are frequently represented as vertex or edge weights. Furthermore, these weights are often not static, instead changing with time and forming a time series. Hence, to fully understand the dynamics of the complex network, we have to consider both network topology and related time series data. In this work, we propose a motif mining approach to identify trend motifs for such purposes. Simply stated, a trend motif describes a recurring subgraph where each of its vertices or edges displays similar dynamics over a userdefined period. Given this, each trend motif occurrence can help reveal significant events in a complex system; frequent trend motifs may aid in uncovering dynamic rules of change for the system, and the distribution of trend motifs may characterize the global dynamics of the system. Here, we have developed efficient mining algorithms to extract trend motifs. Our experimental validation using three disparate empirical datasets, ranging from the stock market, world trade, to a protein interaction network, has demonstrated the efficiency and effectiveness of our approach.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Mon May 28 00:00:00 EDT 2007},
month = {Mon May 28 00:00:00 EDT 2007}
}

Conference:
Other availability
Please see Document Availability for additional information on obtaining the full-text document. Library patrons may search WorldCat to identify libraries that hold this conference proceeding.

Save / Share:
  • We are developing an advanced toolkit for biological networks, using problems from genome biology to drive this work. We now share our experiences in graph analysis and visualization of microbial genome networks using a collection of new and existing graph mining tools and techniques. We address four key problems in genome biology, the organization of complete genome protein networks, feature extraction across chromosomes in microbial strains, the hierarchical structure of protein families and superfamilies, and positional clustering in the genome.
  • Abstract not provided.
  • Computing the set of influential nodes with a given size to ensure maximal spread of influence on a complex network is a challenging problem impacting multiple applications. A rigorous approach to influence maximization involves utilization of optimization routines that comes with a high computational cost. In this work, we propose to exploit the existence of communities in complex networks to accelerate the mining of influential seeds. We provide intuitive reasoning to explain why our approach should be able to provide speedups without significantly degrading the extent of the spread of influence when compared to the case of influence maximization withoutmore » using the community information. Additionally, we have parallelized the complete workflow by leveraging an existing parallel implementation of the Louvain community detection algorithm. We then conduct a series of experiments on a dataset with three representative graphs to first verify our implementation and then demonstrate the speedups. Our method achieves speedups ranging from 3x - 28x for graphs with small number of communities while nearly matching or even exceeding the activation performance on the entire graph. Complexity analysis reveals that dramatic speedups are possible for larger graphs that contain a correspondingly larger number of communities. In addition to the speedups obtained from the utilization of the community structure, scalability results show up to 6.3x speedup on 20 cores relative to the baseline run on 2 cores. Finally, current limitations of the approach are outlined along with the planned next steps.« less
  • The importance of system design in the field of mechanical engineering has been increasingly recognized, but suitable and useful methods for machine design are not yet established. The authors derive easy and useful designing methods for each machine element. They derived a new designing method by linear graph theory and analyzed the results of the dynamic behavior of a power transmission mechanism as an example. The mechanisms which are subject to restrictions of high accuracy and high performance must be designed by considering the characteristic of each machine element. This report deals with a more general power transmission mechanism containingmore » a motor, bearing, and support, and analyzes the results of dynamic behavior. Provided that the analytical model and system graph as in this report are prepared for the mechanism, the performance of the mechanism is easily grasped by this method.« less