Home‎ > ‎

Latest releases

Version 2017-15-01: Major update

posted Jan 15, 2017, 10:44 AM by Mikail Rubinov

New network models

  • generate_fc.m: Generation of synthetic functional connectivity matrices based on structural network measures.
  • predict_fc.m: Prediction of functional connectivity matrices from structural connectivity matrices.
  • mleme_constraint_model.m: Unbiased sampling of networks with soft module and hub constraints (maximum-likelihood estimation of maximum entropy networks).

New measures and demos

  • clique_communities.m: Overlapping community structure via the clique percolation method.
  • rentian_scaling_2d.m and rentian_scaling_3d.m: Updated rentian scaling functions to replace rentian_scaling.m.
  • diffusion_efficiency.m: Global mean and pair-wise effiency based on a diffusion process.
  • distance_wei_floyd.m: All pairs shortest paths via the Floyd-Warshall algorithm.
  • mean_first_passage_time.m: Mean first passage time.
  • path_transitivity.m: Transitivity based on shortest paths.
  • resource_efficiency_bin.m: Resource efficiency and shortest path probability.
  • rout_efficiency.m: Mean, pair-wise and local routing efficiency.
  • retrieve_shortest_path.m: Retrieval of shortest path between source and target nodes.
  • search_information.m: Search information based on shortest paths.
  • demo_efficiency_measures.m: Demonstration of efficiency measures.

Removed functions

  • rentian_scaling.m: Replaced with rentian_scaling_2d.m and rentian_scaling_3d.m.

Bug fixes and/or code improvements and/or documentation improvements

  • efficiency_wei.m: Included a modified weighted variant of the local efficiency.
  • partition_distance.m: Generalized computation of distances to input partition matrices.
  • clustering_coef_wu_sign.m: Fixed computation of the denominator in the Constantini and Perugini versions of the weighted clustering coefficient.
  • modularity_dir.m and modularity_und.m: Updated documentation and simplified code to clarify that these are deterministic algorithms.
  • weight_conversion.m: Corrected bug in weight autofix.

Cosmetic and MATLAB code analyzer (mlint) improvements to many other functions

Version 2016-01-16: Major update

posted Jan 17, 2016, 12:18 PM by Mikail Rubinov   [ updated Jan 17, 2016, 12:23 PM ]

New network models

  • generative_model.m: Implements more than 10 generative network models.
  • evaluate_generative_model.m: Implements and evaluates the accuracy of more than 10 generative network models.
  • demo_generative_models_geometric.m and demo_generative_models_neighbors.m: Demonstrate the capabilities of the new generative model functions.

New network measures

  • clustering_coef_wu_sign.m: Multiple generalizations of the clustering coefficient for networks with positive and negative weights.
  • core_periphery_dir.m: Optimal core structure and core-ness statistic.
  • gateway_coef_sign.m: Gateway coefficient (a variant of the participation coefficient) for networks with positive and negative weights.
  • local_assortativity_sign.m: Local (nodal) assortativity for networks with positive and negative weights.
  • randmio_dir_signed.m: Random directed graph with preserved signed in- and out- degree distribution.

Removed network measures

  • modularity_louvain_und_sign.m, modularity_finetune_und_sign.m: This functionality is now provided by community_louvain.m.
  • modularity_probtune_und_sign.m: Similar functionality is provided by consensus_und.m

Bug fixes and/or code improvements and/or documentation improvements

  • charpath.m: Changed default behavior, such that infinitely long paths (i.e. paths between disconnected nodes) are now included in computations by default, but may be excluded manually.
  • community_louvain.m: Included generalization for negative weights, enforced binary network input for Potts-model Hamiltonian, streamlined code.
  • eigenvector_centrality_und.m: Ensured the use of leading eigenvector for computations of eigenvector centrality.
  • modularity_und.m, modularity_dir.m: Enforced single node moves during fine-tuning step.
  • null_model_und_sign.m and null_model_dir_sign.m: Fixed preservation of negative degrees in sparse networks with negative weights.
  • randmio_und_signed.m: Now allows unbiased exploration of all network configurations.
  • transitivity_bd.m, transitivity_wu.m, transitivity_wd.m: removed tests for absence of nodewise 3-cycles. Expanded documentation.
  • clustering_coef_wu.m, clustering_coef_wd.m: Expanded documentation.
  • motif3-m and motif4-m functions: Expanded documentation.
  • rich_club_wu.m, rich_club_wd.m. Expanded documentation.

Cosmetic and MATLAB code analyzer (mlint) improvements to many other functions

Version 2015-25-01: Major update

posted Jan 17, 2016, 12:16 PM by Mikail Rubinov   [ updated Jan 17, 2016, 12:23 PM ]

Includes two new community-detection scripts and multiple improvements

  • New community detection scripts: 1. community_louvain.m (supersedes modularity_louvain.m and modularity_finetune.m scripts); 2. link_communities.m.
  • added autofix flag to weight_conversion.m for fixing common weight problems.
  • other function improvements: participation_coef.m, charpath.m, reorder_mod.m.
  • bug fixes: modularity_finetune_und_sign.m, modularity_probtune_und_sign.m, threshold_proportional.m
  • changed help files: assortativity_wei.m, distance_wei.m

Version 2014-04-05: Minor update

posted Jan 17, 2016, 12:09 PM by Mikail Rubinov   [ updated Jan 17, 2016, 12:22 PM ]

  • consensus_und.m is now a self-contained function
  • headers in charpath.m and in threshold_proportional.m have been corrected

1-4 of 4