Blog

Updated version of the NOESIS Network Analyzer

posted Aug 7, 2017, 11:32 AM by Fernando Berzal


An updated version of our user-friendly tool for the analysis and visualization of networks, freely available for download at http://noesis.ikor.org/download.



NOESIS Network Analyzer main features

  • Supported network file formats (for analyzing your own networks):

    • GML

    • GraphML

    • GDF

  • Adjustable network visualization:

    • Drag & drop graphical user interface.

    • Automatic layout methods (Fruchterman-Reingold, Kamada-Kawai, hierarchical, radial, random, and regular layouts).

    • Multiple visualization options (styles, colors & sizes).
       
    • Export network images in SVG, PNG, or JPEG format.

  • Network models:

    • Random networks: Erdös-Renyi, Gilbert, Watts-Strogatz, Barabasi-Albert, and Price models.

    • Regular networks: Star, ring, tandem, mesh, toroidal, hypercube, and binary tree networks.

  • Network analysis techniques: 

    • Network structural properties (degree, degree assortativity, eccentricity, average path length, closeness, decay, betweenness, PageRank, HITS, eigenvector centrality, Katz centrality, clustering coefficient, connected components, link betweenness, link embeddedness, link neighborhood overlap...).

    • Community detection methods (Kernighan-Lin partitioning; Newman-Girvan & Radicchi hierarchical community detection; single-link, average-link & complete link hierarchical clustering; fast & multi-step greedy modularity-based community detection; EIG1, KNSC1 & UKMeans spectral community detection, and BigCLAM overlapping community detection).

    • Link scoring & prediction methods (common neighbors, Adamic-Adar score, resource allocation, Jaccard score, preferential attachment, Salton score, Sorensen score, hub-promoted & hub-depressed scores, local & global Leicht-Holme-Newman score, Katz score, random walks & random walks with restarts, flow propagation, pseudoinverse Laplacian score, average commute time score & random forest kernel score).

System requirements: Java Runtime Environment version 8 (JRE8). 

NOTE: The efficient implementation of network analysis techniques makes use of multiple cores in multicore processors when available.



Our Survey of Link Prediction Techniques

posted Jul 3, 2017, 8:58 AM by Fernando Berzal   [ updated Jul 3, 2017, 9:02 AM ]

As part of the Ph.D. work by Víctor Martínez, he performed an extensive bibliographic study of link prediction techniques and, as far as we know, the most extensive empirical evaluation of link prediction techniques that is currently available.


Our survey has been published by ACM Computing Surveys, one of the top journals in the field:

Víctor Martínez, Fernando Berzal & Juan-Carlos Cubero:
"A Survey of Link Prediction in Complex Networks"
ACM Computing Surveys, Volume 49, Issue 4, Article No. 69, February 2017 


According to Thomson Reuters Journal Citation Reports, in 2016, ACM Computing Surveys had an impact factor of 6.748, being the second top-ranked journal in the "Computer Science - Theory & Methods" category (i.e. 2nd out of 104 research journals). The JCR impact factor is a measure reflecting the yearly average number of citations to recent articles published in that journal. It is frequently used as a proxy for the relative importance of a journal within its field. Journals with higher impact factors are often deemed to be more important than those with lower ones. The impact factor was devised by Eugene Garfield, the founder of the Institute for Scientific Information.

Given that ACM Computing Surveys published 143 papers in the last two years (104 papers in 2015 and 39 papers in 2014) and those papers received 965 citations in 2016 (386 citations to papers published in 2015 and 579 citations to papers published in 2014), its impact factor is just the result of dividing the number of citations to recent items (965) by the number of recently-published papers (143): 965/143 = 6.748.

[Project status] Software metrics

posted Jun 22, 2017, 4:17 AM by Fernando Berzal   [ updated Jun 22, 2017, 4:21 AM ]

The following tables collect some software metrics corresponding to the two main components of the NOESIS framework: the NOESIS core classes and interfaces, which implement a wide range of network analysis algorithms, and the reusable iKor component library, which provides parallelization support, mathematical routines, a collection framework, and a model-driven application generator for the NOESIS GUI.

The metrics were obtained using Campwood Software's excellent SourceMonitor utility, available at http://www.campwoodsw.com/sourcemonitor.html

NOESIS Network Data Mining Framework

Date: June 22nd, 2017
Source code repository: https://bitbucket.org/fberzal/noesis

MetricProductionTestTotal
Files
29281373
Lines26400886135261
Statements12184474316927
% Branch statements12.4%3.2%9.8%
% Lines with comments11.9%3.2%9.7%
Classes and interfaces35584439
Methods per class3.884.574.01
Avg. Statements per method5.628.906.34
Avg. Block depth1.841.621.78
Avg. Complexity2.211.422.04


iKor Reusable Component Library

Date: June 22nd, 2017
Source code repository: https://bitbucket.org/fberzal/ikor

MetricProductionTestTotal
Files
29187378
Lines294791302342502
Statements11339661717956
% Branch statements12.5%2.7%8.9%
% Lines with comments17.2%6.0%13.8%
Classes and interfaces31990409
Methods per class6.248.086.64
Avg. Statements per method3.686.824.52
Avg. Block depth1.891.671.81
Avg. Complexity1.821.241.66


Link scoring and prediction methods

posted May 15, 2015, 1:47 AM by Fernando Berzal   [ updated Jun 22, 2017, 3:17 AM by Fernando Berzal ]

A version of the NOESIS Network Analyzer tool including a score of link scoring and prediction methods is now available here.

NOESIS Network Analyzer

The tool now supports the following methods, which can be used both for link scoring (i.e. evaluating existing link) and link prediction (i.e. predicting new links):

- CN (Common Neighbors).

- AA (Adamic-Adar score).

- RA (Resource Allocation score).

- J (Jaccard score).

- PA (Preferential Attachment score).

- Salton score.

- Sorensen score.

- HPI (Hub-promoted index)

- HDI (Hub-depresseed index).

- LLHN (Local Leicht-Holme-Newman score)

- GLHN (Global Leicht-Holme-Newman score)

- K (Katz score)

- RW (random walk score).

- RWR (random walk with restart score).

- FP (flow propagation score).

- PL (pseudoinverse Laplacian score).

- ACT (average commute time score).

- RFK (random forest kernel score).


      
Victor Hugo's "Les Misérables" link scoring: common neighbours, preferential attachment score & Katz index.


You can access these methods using the "Analysis > Links > Link prediction" menu for predicting new links or the "Analysis > Links > Link scoring" menu for evaluating existing ones. They are available in the LSP [link scoring and prediction] version of the NOESIS Network Analyzer.

NOESIS Network Analyzer

posted Oct 22, 2014, 3:51 AM by Fernando Berzal   [ updated Nov 8, 2015, 11:35 PM ]

An user-friendly tool for the analysis and visualization of networks, freely available for download at http://noesis.ikor.org/download.

http://goo.gl/T0juNB


NOESIS Network Analyzer main features

  • Supported network file formats (for analyzing your own networks):

    • GML

    • GraphML

    • GDF

  • Adjustable network visualization:

    • Drag & drop graphical user interface.

    • Automatic layout methods (Fruchterman-Reingold, Kamada-Kawai, hierarchical, radial, random, and regular layouts).

    • Multiple visualization options (styles, colors & sizes).
       
    • Export network images in SVG, PNG, or JPEG format.

  • Network models:

    • Random networks: Erdös-Renyi, Gilbert, Watts-Strogatz, Barabasi-Albert, and Price models.

    • Regular networks: Star, ring, tandem, mesh, toroidal, hypercube, and binary tree networks.

  • Network analysis techniques: 

    • Network structural properties (degree, degree assortativity, eccentricity, average path length, closeness, decay, betweenness, PageRank, HITS, eigenvector centrality, Katz centrality, clustering coefficient, connected components, link betweenness, link embeddedness, link neighborhood overlap...).

    • Community detection methods (Kernighan-Lin, Newman-Girvan, Radicchi, hierarchical, fast greedy, spectral, and BigCLAM community detection).

System requirements: Java Runtime Environment version 8 (JRE8). 

NOTE: The efficient implementation of network analysis techniques makes use of multiple cores in multicore processors when available.


http://goo.gl/1wRzOt

http://goo.gl/1wRzOt



NOESIS Network Viewer

posted Mar 12, 2014, 11:21 AM by Fernando Berzal   [ updated Mar 12, 2014, 11:36 AM ]

An easy-to-use tool for the analysis and visualization of networks, freely available for download...

http://noesis.ikor.org/download


Main features

  • Supported network file formats: GML, GraphML, and GDF.

  • Export images in SVG, PNG, or JPEG format.

  • Random network models: Erdös-Renyi, Gilbert, Watts-Strogatz, Barabasi-Albert, and Price models.

  • Network analysis: degree, eccentricity, average path length, closeness, decay, betweenness, PageRank, HITS, eigenvector centrality, Katz centrality, clustering coefficient, connected components, link betweenness, link embeddedness, link neighborhood overlap...

Network examples


Gnutella P2P network


International trade network



Random scale-free networks (Barabasi-Albert model)

Contrato predoctoral / Ph.D. position

posted Aug 27, 2013, 3:11 AM by Fernando Berzal   [ updated Aug 27, 2013, 3:14 AM ]

Contrato predoctoral asociado al proyecto de investigación NOESIS

Proyecto TIN2012-36951
"NOESIS: Network-Oriented Exploration, Simulation, and Induction System"


Convocatoria: BOE del 14 de agosto de 2013,
https://www.boe.es/diario_boe/txt.php?id=BOE-A-2013-8984

Plazo de solicitud: Hasta el 10 de septiembre de 2013.

Duración del contrato: 4 años.

Condiciones del contrato: Contrato de 20600 euros al año,con una retribución salarial mínima de 16422 euros brutos y una ayuda de 1500 euros para financiar el pago de la matrícula de doctorado.

Requisitos: Haber finalizado los estudios que permitan acceder a un programa de doctorado después del 1 de enero de 2010.

Página web del programa de ayudas para contratos predoctorales para la formación de doctores 2013 (Secretaría de Estado de I+D+i, Ministerio
de Economía y Competitividad): http://goo.gl/nxT2UE


Información adicional:

  Fernando Berzal (berzal@acm.org)
  Departamento de Ciencias de la Computación e Inteligencia Artificial
  Universidad de Granada

Parallelism @ NOESIS

posted Jan 10, 2013, 11:06 AM by Fernando Berzal   [ updated Jan 11, 2013, 3:20 PM ]

Simple performance test using a Wikipedia GML network (available at http://spark-public.s3.amazonaws.com/sna/other/wikipedia.gml):

NETWORK STATISTICS
- Nodes: 27475
- Links: 85729
- 3 node attributes: id wikiid label
- 0 link attributes:
Degree distributions
- Out-degrees: [n=27475 min=0.0 max=565.0 avg=3.1202547770700635 dev=9.038219683086334]
- In-degrees:  [n=27475 min=0.0 max=367.0 avg=3.1202547770700635 dev=8.99990229087909]
Node of maximum out-degree:
- out-degree: 565 out-links
- in-degree: 0 in-links
- id: 8436
- wikiid: 1807178
- label: List of mathematics articles (S)
Node of maximum in-degree:
- out-degree: 44 out-links
- in-degree: 367 in-links
- id: 10807
- wikiid: 7250299
- label: Geometry
Betweenness
[n=27475 min=2.0 max=2.1696583120297994E7 avg=79690.76047315753 dev=404883.4016065179]
Node of maximum betweenness:
- out-degree: 29 out-links
- in-degree: 82 in-links
- id: 12533
- wikiid: 5176
- label: Calculus

Time:  27442 ms

Using a conventional Core i5 laptop (vs. 56 seconds using a sequential implementation). 

Same experiment running igraph... 103 seconds 

Same experiment running NetworkX... 40 minutes !!!


UPDATE: 
Additional tests on a 2.67GHz Intel Core i7 920 desktop PC using different CPU schedulers:
- 71.6-71.9s @ SequentialScheduler
- 19.6-26.7s @ FutureScheduler(8)
- 16.7-20.1s @ WorkStealingScheduler(8)
- 16.7-19.4s @ FutureScheduler(32)
- 16.6-19.3s @ FutureScheduler(16)
- 16.5-17.8s @ ThreadPoolScheduler
- 16.4-17.1s @ WorkStealingScheduler(64)
- 16.4-17.0s @ WorkStealingScheduler(32)
- 16.4-16.6s @ WorkStealingScheduler(16)

[SW Metrics] iKor Library

posted Nov 1, 2012, 4:20 AM by Fernando Berzal   [ updated Nov 1, 2012, 4:20 AM ]

Project: iKor Library
Checkpoint: r312 (SkipArray ADT)
Date: November 1st, 2012

MetricProductionTestTotal
Files
7726103
Lines9398390513303
Statements350521305635
% Branch statements17.3%2.0%11.6%
% Lines with comments22.3%7.5%17.9%
Classes and interfaces8027107
Methods per class8.645.857.93
Avg. Statements per method3.4010.024.63
Avg. Block depth1.801.661.75
Avg. Complexity2.081.261.91

[SW Metrics] NOESIS

posted Oct 14, 2012, 4:00 AM by Fernando Berzal

Project: NOESIS
Checkpoint: r305 (network readers & writers for GDF, GML & GraphML file formats)
Date: October 14th, 2012

MetricProductionTestTotal
Files
6634100
Lines6621396910590
Statements294221755117
% Branch statements16.2%2.6%10.4%
% Lines with comments8.3%1.8%5.8%
Classes and interfaces8138119
Methods per class5.744.685.40
Avg. Statements per method4.328.755.55
Avg. Block depth1.891.541.75
Avg. Complexity2.121.331.91

1-10 of 14