https://visone.info/wiki/index.php?title=Special:NewPages&feed=atom&hideredirs=1&limit=50&offset=&namespace=0&username=&tagfilter=&size-mode=max&size=0visone manual - New pages [en]2021-11-28T20:14:48ZFrom visone manualMediaWiki 1.32.0https://visone.info/wiki/index.php/POLNET_(data)POLNET (data)2018-06-18T10:59:10Z<p>Lerner: </p>
<hr />
<div>The following ZIP file contains exemplary data used in introductory visone workshops. <br />
<br />
[[Media:Visone_data.zip|'''Visone_data.zip''']] (1.36 MB)<br />
<br />
Note that visone cannot directly load data from ZIP files. So you should extract (''unzip'') the files on your computer before opening them in visone.<br />
<br />
In addition, you should [http://visone.info/html/download.html '''download'''] the current version of visone and (as a test) execute it as it is described in the very first sentence of the [[Installation_(tutorial)|installation tutorial]].<br />
<br />
The following online tutorials repeat or extend content of the visone session at POLNET+<br />
<br />
* [[Visualization_and_analysis_(tutorial)|'''Visualization and analysis''']]<br />
* [[Personal_networks_(tutorial)|'''Personal networks''']]<br />
* [[Collections_(tutorial)|'''Collections''']]</div>Lernerhttps://visone.info/wiki/index.php/Quad_censusQuad census2017-08-03T14:06:03Z<p>Ortmann: </p>
<hr />
<div>=Quad Census=<br />
===Method===<br />
[[File:Quad_census.png|400px|thumb|All non-isomorphic subgraphs with four nodes (quads). Node and edge labels refer to the orbits and were enumerated such that each orbit is identified with a single quad.]]<br />
<br />
The frequency distribution over all non-isomorphic 4-node subgraphs is called the quad census.<br />
<br />
The quad census analysis calculates for each node and edge respectively how often it is contained in a specific quad. Depending on the chosen parameters this analysis further distinguishes between different orbits. For example a claw has two node orbits, i.e., 11 and 12. Considering the orbits leads to a finer grained description of the nodes/edges and allows for a more precise comparison. <br />
<br />
=== Complexity ===<br />
The computation of the orbit-aware Quad Census can be done in <math>O(a(G)^2m)</math> where <math>m</math> denotes the number of edges of the graph and <math>a(G)</math> is the arboricity of the graph.<br />
<br />
More detailed information on the computation of the orbit-aware Quad Census is available in<br />
* Mark Ortmann, and Ulrik Brandes: [https://link.springer.com/content/pdf/10.1007%2Fs41109-017-0027-2.pdf Efficient orbit-aware triad and quad census in directed and undirected graphs], Applied Network Science, 2(13), 2017.<br />
<br />
=== Settings ===<br />
* orbit-aware frequencies: If checked the node and edge orbit frequencies are computed, otherwise orbits will not be distinguished.<br />
* write non-induced frequencies: If checked besides the induced also the node and edge (orbit-aware) non-induced frequencies are written.<br />
* write network frequencies: Additionally computes the quad census on a graph level (if checked)<br />
<br />
=== Result ===<br />
The result of this analysis are written to node/edge and network attributes:<br />
* ind_orbit_..: The induced orbit-aware frequencies. The orbit number corresponds to the numbering shown in the opposing figure<br />
* non-ind_orbit_...: Same as ind_orbit_.. but for non-induced frequencies<br />
* <quad_name>: If the orbit-aware frequencies box is not checked the names of the corresponding quads is used, e.g. claw or diamond</div>Ortmannhttps://visone.info/wiki/index.php/Sparse_stress_minimizationSparse stress minimization2017-02-22T11:34:59Z<p>Ortmann: </p>
<hr />
<div>=Sparse Stress Layout=<br />
===Method===<br />
The sparse stress layout is a heuristic to approximate the (full) stress layout. The goal of the stress minimization is to find a layout, <math>x</math>, such that <math>\sum_{i<j}w_{ij}(||x_i - x_j|| - d_{ij})^2 \text{ is as small as possible.}</math> In other words stress minimization tries to find a layout in which the euclidean distance of each pair of nodes matches their graph-theoretical distance, i.e. shortest-path distance. <br />
<br />
In comparison to the (full) stress minimization, sparse stress on the one hand requires less space and running time, but on the other hand the resulting layouts have a slightly lower quality. The space and running time reduction is achieved by restricting the stress function from <math>n^2</math> to <math>m+np</math> with <math>p</math> denoting the number of pivots. <br />
<br />
More detailed background information can be found in<br />
* Mark Ortmann, Mirza Klimenta, and Ulrik Brandes: [https://dx.doi.org/10.1007/978-3-319-50106-2_2 A sparse Stress Model], GD'16, 18-32, 2016.<br />
<br />
=== Complexity ===<br />
The computation of the sparse stress layout requires <math>O(m+np)</math> time and space and a <math>O(p(m+n \log n))</math> preprocessing time.<br />
<br />
=== Parameters & Their Influence ===<br />
* By its iterative nature the quality of the stress minimized layout depends on the initialization. Therefore, we recommend to check the "init with PivotMDS" box.<br />
* The number of pivots influences the quality of the final layout. A higher number of pivots increases the quality, but also space and running time.<br />
* The pivot strategy allows to choose between various (pivot) sampling strategies. We recommend K_MEANS_SSSP as it has advantages over the other routines respective the layout quality. However if time is the most important factor we recommend Random. The other two strategies are a compromise between quality and running time. <br />
* The link lengths defines the distance between adjacent nodes in the euclidean space and is used for the shortest-path computation.<br />
* The maximum number of iterations defines after how many executions of the iterative stress minimization algorithm the procedure should stop. A higher number of iterations results in a higher quality.<br />
* Given the case that the layouts of two consecutive iterations of the sparse stress algorithm differ only by a very small amount we say that the algorithm has converged. Since, often the number of iterations is often higher than necessary, we recommend to check the "stop on convergence" box to decrease the running time of the algorithm.</div>Nocajhttps://visone.info/wiki/index.php/Edge_ConnectivityEdge Connectivity2015-08-07T13:54:49Z<p>Nocaj: Created page with "The edge (or link) connectivity between two nodes u and v is the number of edges which have to be removed such that u and v are not connected by a path anymore."</p>
<hr />
<div>The edge (or link) connectivity between two nodes u and v is the number of edges which have to be removed such that u and v are not connected by a path anymore.</div>Nocajhttps://visone.info/wiki/index.php/Node_ConnectivityNode Connectivity2015-08-07T13:40:28Z<p>Nocaj: </p>
<hr />
<div>The node connectivity between two nodes u and v is the number of nodes which have to be removed such that u and v are not connected by a path anymore. <br />
Link direction is not considered.<br />
<br />
If u and v are connected by a direct link (u,v) then their node connectivity is the number of nodes in their connected component minus one.</div>Nocajhttps://visone.info/wiki/index.php/Transitive_ReductionTransitive Reduction2015-07-29T13:32:47Z<p>Nocaj: </p>
<hr />
<div>In a directed acyclic graph, every transitive link is removed. This transformation maintains the reachability among the nodes.<br />
<br />
If the graph has cycles every transitive link between two different strongly connected components is removed.<br />
<br />
See https://en.wikipedia.org/wiki/Transitive_reduction for a more detailed discussion.</div>Nocajhttps://visone.info/wiki/index.php/Effective_ResistanceEffective Resistance2015-07-09T06:26:55Z<p>Nocaj: Created page with "Computes the pairwise effective resistance (also called resistance distance) for all pairs of nodes. See the following article as a detailed reference: * Klein, D. J., & RandiÄ‡,..."</p>
<hr />
<div>Computes the pairwise effective resistance (also called resistance distance) for all pairs of nodes.<br />
See the following article as a detailed reference:<br />
* Klein, D. J., & RandiÄ‡, M. (1993). Resistance distance. Journal of Mathematical Chemistry, 12(1), 81-95.</div>Nocajhttps://visone.info/wiki/index.php/Dynamic_layoutDynamic layout2015-06-14T12:50:22Z<p>Nataschah: </p>
<hr />
<div>visone extends the visualization of networks at each moment in time to the animation of [[network collection|networks collections]] '''over time'''. Such animation is showing a movie that takes you from the first network to the last one of an active network collection. Therefore a '''dynamic layout''' is needed: a dynamic layout algorithm computes a new layout for each network in the collection that enables a meaningful animation. visone offers three dynamic layouts, each with a different focus:<br />
<br />
* '''aggregated layout''' computes the aggregation network, that is, the ''union'' of all networks in the collection. The aggregation network has as nodes all nodes that are present in at least one of the networks with links connecting nodes that are adjacent at some time point. This aggregation network is layouted once, defining the nodes' positions for all time points. The resulting animation, thus, has a maximum stability (node do not move at all) and just shows which ties emerge or disappear over time. <br />
* '''anchored layout''' The anchored layout allows to specify a tradeoff between stability and optimal network layout at the individual time points. The nodes' positions for a single time point are computed by optimizing the quality function of stress minimization requiring at the same time that nodes are not too far from their positions in the "anchor network". (The network can be anchored at the aggregation network or at the network from the preceeding time point.) If the '''stability''' parameter is set to a high value, nodes move only slighly from one time point to the next; a low '''stability''' means that layouts at individual time points are computed paying little attention to the anchor network. <br />
* '''linked layout''' ...<br />
<br />
<br />
To compute a [[dynamic layout]] go to the [[visualization tab]] and chose in the drop-down menus (from top to bottom) ''layout'', ''node layout'', and ''dynamic layout''. Choose one of the three dynamic layout methods and click on the "layout collection" button. After computing the dynamic layout, click on the [[File:Zoom_fitall.png|link=view_menu#fit_all_networks]] icon in visone's [[GUI#toolbar|toolbar]] to fit all networks into a common bounding box. To open the animation window click on the [[File:Animation.png|link=animation]] icon in visone's [[GUI#toolbar|toolbar]] and click on the "play" button to start the animation. For an exhaustive explanation of dealing with network collections and dynamic visualization see the [[Collections (tutorial)|network collections tutorial]].</div>Nataschahhttps://visone.info/wiki/index.php/DAG_Longest_PathDAG Longest Path2015-06-10T12:57:52Z<p>Nocaj: </p>
<hr />
<div>Each node gets the length of the longest directed path from any source in the directed acyclic graph (DAG).<br />
A source in DAG is a node which has only outgoing links.<br />
<br />
If the graph has cycles, each strongly connected component is aggregated to one node which results in a DAG.<br />
In this case, each vertex gets the longest directed path from any strongly connected component whose nodes do not have any incoming edge to a node of another strongly connected component.</div>Nocajhttps://visone.info/wiki/index.php/Loop_To_Node_AttributeLoop To Node Attribute2015-06-10T11:56:59Z<p>Nocaj: moved Loop Attribute To Node Attribute to Loop To Node Attribute over redirect</p>
<hr />
<div>For each selected edge attribute a node attribute is created containing the value of the self loop.<br />
For multiple self loops aggregation is performed based on the selected action.</div>Nocajhttps://visone.info/wiki/index.php/Triangle_k_CoreTriangle k Core2015-06-08T14:20:41Z<p>Nocaj: </p>
<hr />
<div>The triangle k core, or k truss has been intoduced by [http://www.csee.ogi.edu/~zak/cs506-pslc/trusses.pdf Cohen (NSA Tech. Report 2008)] and [http://web.cse.ohio-state.edu/~zhangya/ICDE12_conf_full_179.pdf Zhang and Parthasarathy (ICDE 2012)] independently and runs in <math>\mathcal{O}(\Delta(G)m)</math> time, where <math>\Delta(G)</math> is the maximum degree and <math>m</math> the number of edges.<br />
<br />
=== Definition Triangle Core ===<br />
The triangle k-core of a simple undirected graph <math>G = (V,E)</math> is the inclusion maximal subgraph <math>C_{k}(G) \subset G</math> where each edge <math>e \in E(C_k(G))</math> is part of at least <math>k</math> triangles in <math>C_k(G)</math>.<br />
<br />
More detailed background information is provided in<br />
* Cohen, Jonathan. "Trusses: Cohesive subgraphs for social network analysis." National Security Agency Technical Report (2008).<br />
* Zhang, Yang, and Srinivasan Parthasarathy. "Extracting analyzing and visualizing triangle k-core motifs within networks." In Data Engineering (ICDE), 2012 IEEE 28th International Conference on, pp. 1049-1060. IEEE, 2012.<br />
<br />
=== Definition Triangle Core Number ===<br />
The triangle core number of an edge <math>e \in E(G)</math> is the maximal k such that <math>e \in C_k(G)</math></div>Nocajhttps://visone.info/wiki/index.php/Triangle_k_coreTriangle k core2015-06-08T14:20:10Z<p>Nocaj: Created page with "will be filled soon..."</p>
<hr />
<div>will be filled soon...</div>Nocajhttps://visone.info/wiki/index.php/Positional_DominancePositional Dominance2015-06-08T13:38:32Z<p>Brandes: sketch of functionality</p>
<hr />
<div>Positional dominance is a powerful generic concept that unifies the majority of centrality indices and role equivalences.<br />
It defines a relation on the nodes of a network, where one node dominates another if it has stronger relationships with comparable others.<br />
The result is a new network with the same set of nodes and directed edges <math>(i,j)</math> whenever the position of node <math>i</math> dominates the position of node <math>j</math> in the network.<br />
<br />
The position of a node is defined by either the incoming or outgoing links of that node.<br />
The strengths of links may vary with a numeric link attribute.<br />
When comparing two positions, we are really comparing relationships.<br />
A relationship is stronger than another, if its link attribute is at least as large ''and'' both source and target are comparable to the source and target of the other relationship.<br />
The position of a node dominates that of another, if the relationships defining it are stronger.<br />
<br />
The simplest example is the following: <br />
* uniform link strength<br />
* outgoing links<br />
* exclude compared dyad<br />
* source comparability: grouped by uniform (no a-priori distinctions between nodes, all sources created equal) <br />
* target comparability: unique (the relationship to a target can only be matched by a relationship to the very same target)<br />
<br />
This results in a network in which a node dominates another positionally if it has outgoing links to the very same other nodes (and, possibly, more).<br />
Now include the compared dyad and switch target comparability to ''grouped by uniform''; with these new settings, a node dominates another positionally, if it's outdegree is no less.<br />
Finally, by checking ''single-link dominance'', we can have a node dominate any other node simply by having a non-zero degree.<br />
<br />
'''NB:''' Currently, the provision of positional dominance serves as a preview to future forms of network analysis.<br />
A comprehensive explanation of the utility of network positions is in preparation.</div>Nocajhttps://visone.info/wiki/index.php/Complement_GraphComplement Graph2015-06-08T10:18:26Z<p>Mader: Created page with "This function creates the complement graph corresponding to the current network. The resulting complement graph contains a link between two nodes exactly if there is no link in t..."</p>
<hr />
<div>This function creates the complement graph corresponding to the current network. The resulting complement graph contains a link between two nodes exactly if there is no link in the original graph and vice versa.</div>Maderhttps://visone.info/wiki/index.php/Louvain_ClusteringLouvain Clustering2015-03-12T13:42:50Z<p>Nocaj: </p>
<hr />
<div>=Louvain Clustering=<br />
=== Method ===<br />
<br />
The Louvain clustering tries to optimize modularity in a greedy fashion by randomly moving nodes from one cluster to another in multiple levels.<br />
<br />
The algorithm is:<br />
# (level) start with each node being a singleton cluster:<br />
# consider nodes in random order<br />
# iterate as long as cluster membership changes<br />
#* for each node : remove it from its current cluster and add it to the cluster with the highest modularity gain<br />
# aggregate the resulting clustering to a new graph and continue on next level (step 1), as long as modularity improves.<br />
<br />
<br />
=== Complexity ===<br />
Practically, the algorithm seems to scale well for large graphs.<br />
<br />
=== References ===<br />
*Fast unfolding of communities in large networks, Vincent D Blondel, Jean-Loup Guillaume, Renaud Lambiotte, Etienne Lefebvre, Journal of Statistical Mechanics: Theory and Experiment 2008 (10), P10008 (12pp) doi: 10.1088/1742-5468/2008/10/P10008. ArXiv: http://arxiv.org/abs/0803.0476<br />
*"The Louvain method for community detection in large networks" Vincent Blondel http://perso.uclouvain.be/vincent.blondel/research/louvain.html</div>Nocajhttps://visone.info/wiki/index.php/Geographic_MappingGeographic Mapping2015-03-12T12:43:14Z<p>Nocaj: </p>
<hr />
<div>In geographical networks the layout of nodes can be determined by some geographic coordinates.<br />
<br />
The geographic mapping allows the representation of geographical networks, whose node positions are characterized by longitude and latitude, over the corresponding map. <br />
<br />
The main features are:<br />
* map nodes: positions based on two node attributes (longitude latitude) using mercator projection<br />
* update map: download a map from the selected open street map provider and map it on the background (internet connection required)<br />
<br />
If just the nodes should be mapped, e.g., without redownloading a new map, then "update map" can be deactivated.<br />
<br />
Several map styles and different levels of detail can be selected by the user.<br />
<br />
<br />
{|<br />
|+ valign="top"| Communication network: <br />
| 9 students at the University of Konstanz living in the Bodensee region and communicating during the semester break<br />
|- halign="center"<br />
| style="text-align: center;" |[[File:geomap.png|400px]]<br />
|<br />
|}<br />
<br />
=== Example Data ===<br />
The communication network of 9 students at the University of Konstanz during the semester break:<br />
* graphml file with network containing geographic node attributes [[File:geonet-withoutMap.graphmlz]]<br />
* graphml file with geographically mapped nodes and map [[File:geonet-withMap.graphmlz]]</div>Nocaj