Skip to content
SocNetV SocNetV
News

Bug Fixes

3 posts with the tag “Bug Fixes”

SocNetV v1.9 released - Bug Fixes and Speed Increase

Dimitris

The Social Network Visualizer project has just released version 1.9, which fixes many important bugs and brings a faster matrix inverse routine. Source code, Windows zipped executables, Mac OS image, and binary packages for major Linux distributions are as always available from the Downloads page.

Key Improvements

  • Faster Matrix Inverse Routine:
    The matrix inverse algorithm now uses LU decomposition, greatly improving computation speed. This enhancement also affects the Information Centrality algorithm, which now runs in 1/10th of the time required in earlier SocNetV versions.

  • Revamped PageRank Prestige Algorithm:
    Up to version 1.8, the PageRank algorithm used the original Page & Brin formula, leading to different results. Starting from this version, SocNetV uses the correct formula and computes comparable results. Additionally, the initial PageRank score of each node is now set to 1/N.

Closed Bugs in Version 1.9

  • #1463069: Wrong average distance when there are isolates.
  • #1365037: Certain sparse matrices crash SocNetV on the invertMatrix method.
  • #1365582: CentralityInformation() is slow when network size (N > 100).
  • #1463095: Edge filter works but the user cannot undo.
  • #1464422: Incorrect PageRank results.
  • #1464430: SocNetV refuses to read Pajek files not starting with *Network.
  • #1465774: Edges do not always follow relations.
  • #1463082: Edge color change does not take effect.
  • #1464418: SocNetV crashes during PageRank computation on isolated nodes.

With these enhancements and fixes, SocNetV v1.9 ensures faster and more reliable performance for your social network analysis tasks.

SocNetV 1.5 Released

Dimitris

A new version of SocNetV has just been released! Version 1.5 brings standardization, many bug fixes, and exciting new features. Packages for Windows, Mac OS X, and Linux are available from the Downloads page.

Notable New Features

  • Prominence Scores on Valued Networks:

    • A new SSSP-solver algorithm (Dijkstra) enables prominence indices on weighted networks.
    • SocNetV asks whether to consider or invert weights, depending on their meaning (e.g., cost vs. votes).

  • Standardized Centrality and Prestige Scores:

    • Scores are now standardized (0.0 to 1.0) where applicable.
    • If no standardized formula exists, SocNetV normalizes by dividing the original score by the sum of all scores.

  • Prominence Layouts Relative to Highest Score:

    • Visualization layouts (circular, level, and nodal size) are now based on the highest score in the network rather than theoretical maximums.

  • Skip Isolates:

    • Option to omit isolates when computing prominence indices.

  • Graph Connectedness:

    • SocNetV reports if the network is connected or has isolates and suggests fixes.

  • CC Drops Isolates by Default:

    • Automatically drops isolate nodes to compute CC scores, avoiding user interruptions.

  • New Datasets:

    • Stephenson and Zelen (1989): Network of 40 AIDS patients.
    • Stephenson and Zelen (1989): IC test dataset (5 actors).
    • Wasserman and Faust: Star, circle, and line graphs (7 actors).

Bug Fixes

The following bugs have been resolved in this release:

  • #1358678: Fix GDC calculation in weighted networks.
  • #379558: Force-directed algorithms produce poor layouts.
  • #1365028: Methods isOutLinked & setOutLinked do not consider relations.
  • #1365504: centralityInformation() should symmetrize the adjacency matrix.
  • #1366625: Clicking on an edge does not select the correct edge.
  • #1369171: Group IC calculation yields incorrect results.
  • #1371208: Incorrect power centrality scores.
  • #1364955: vertices() should report only enabled vertices.
  • #1369336: Pagerank prestige reports incorrect scores.
  • #1370528: SocNetV cannot build on non-x86 architectures.
  • #1364320: Remove SRS PDF and fix spelling errors in code.
  • #1364361: SocNetV does not remember the last directory used by the user.
  • #1378346: Cannot change size and value of a node.

Enjoy the new features and improvements! If you spot any bugs, please report them here. Have fun!

SocNetV v1.4 Released!

Dimitris

Today is a fine day to release another SocNetV update. Version 1.4 brings consistency to the application by fixing many long-standing bugs (e.g., #514264 and #713617) and adding exciting new features.

New Features

  • Node Size Layouts Based on Prominence Scores:

    • SocNetV can now adjust node sizes to reflect a selected prominence index score.
    • This feature supports all indices calculated by SocNetV, including Degree, Closeness, Influence Range Closeness, Betweenness, PageRank, Proximity, Eccentricity, and Power.

  • Support for UCINET Edgelist1 Format:

    • SocNetV can now import edgelist1 UCINET format, for example:
dl
N=48
format=edgelist1
data:
1 2 4
1 3 2
1 6 2
1 8 2

  • Updated Shortcuts for Node and Edge Removal:

    • Node removal: Ctrl+Backspace
    • Edge removal: Shift+Backspace
    • You can also click on a node and press Ctrl+Backspace to remove it.

  • New Dataset:

    • Freeman’s EIES networks are now included:
      • Multirelational mode: 32 actors.
      • Single relation mode: 48 actors.

Bug Fixes

This release addresses many longstanding issues, enhancing stability and usability. See the complete ChangeLog for more details.

Downloads

Source code, packages, and executables for Linux, Mac OS X, and Windows are available on the Downloads page.

Have fun!