Component Ratio calculation to a node adjacent matrix #help
|
Hi all,
I have the directional networks from 1990 to 2020 - the files attached belongs to 2020, and I'm trying to calculate the component ratio of each node adjacent matrix, icluding the ties among adjacent nodes. Does any one have an idea how to generate automatically the sub-digraphs for the calculation? Saluti? matui |
¿ªÔÆÌåÓýPaulo, the component ratio is something you calculate for the whole network. It measures the extent to which the network is broken up into fragments. In UCINET you can calculate it by going to network|whole networks|multiple measures. For your directed data, you need to decide whether to respect direction or ignore it. I ran it both ways and got the below. First column is ignoring direction and the second is respecting direction. ? ? ?????????????????????????????????? 1????????? 2 ??????????????????????????000_2020_C 000_2020_C ??????????????????????????ONS_ACAD_t ONS_ACAD_t ??????????????????????????o_PROF-coh o_PROF-coh ???????????????????????????????????u????????? d ??????????????????????????---------- ---------- ????1???????? # of nodes???? 317.000??? 317.000 ????2????????? # of ties???? 740.000??? 416.000 ????3???????? Avg Degree?????? 2.334????? 1.312 ????4????? Indeg H-Index?????? 7.000????? 5.000 ????5?????? K-core index?????? 4.000????? 4.000 ????6 Deg Centralization?????? 0.075????? 0.075 ????7 Out-Centralization?????? 0.075????? 0.021 ????8? In-Centralization?????? 0.075????? 0.066 ????9???????? Indeg Corr?????? 0.062????? 0.056 ???10??????? Outdeg Corr?????? 0.062????? 0.027 ???11??????????? Density?????? 0.007????? 0.004 ???12???????? Components????? 54.000??? 241.000 ???13??? Component Ratio?????? 0.168????? 0.759 ???14????? Connectedness?????? 0.513????? 0.062 ???15????? Fragmentation?????? 0.487????? 0.938 ???16??????????? Closure?????? 0.186????? 0.135 ???17?????? Avg Distance?????? 6.021????? 4.506 ???18????? Prop within 3?????? 0.074????? 0.019 ???19?????????? # w/in 3??? 7390.000?? 1861.000 ???20??????? SD Distance?????? 2.408????? 1.884 ???21?????????? Diameter????? 16.000???? 11.000 ???22?????? Wiener Index? 309716.000? 28206.000 ???23???? Dependency Sum? 258280.000? 21947.000 ???24??????????? Breadth?????? 0.894????? 0.982 ???25??????? Compactness?????? 0.106????? 0.018 ???26??? Small Worldness??????????????????????? ???27??????????? Mutuals?????? 0.007????? 0.001 ???28??????? Asymmetrics?????? 0.000????? 0.006 ???29????? ????????Nulls?????? 0.993????? 0.993 ???30??? Arc Reciprocity?????? 1.000????? 0.221 ???31?? Dyad Reciprocity?????? 1.000????? 0.124 ? Having looked at your data, I¡¯m not sure the component ratio would be very useful. If we ignore direction, your data consist of a single large component containing 72% of all nodes, and many many tiny components that are just isolates or dyads. ? ? The picture seems more valuable than the statistic. ? steve ? ? From: [email protected] <[email protected]> On Behalf Of Paulo Matui
Sent: Tuesday, June 7, 2022 15:01 To: [email protected] Subject: [ucinet] Component Ratio calculation to a node adjacent matrix #help ? Hi all, |
|
Hi Professor, I know, I?did this analysis?with this metric in the figure attached. Actually, reading my?research notes I have a question about the CR metric. It is the number of components minus 1 divided?by the number of nodes minus 1. To achieve the number of strong components in a digraph, does UCINET?use the in-depth research from Robert Tarjan (1972)? This is the population of Administration and Industrial Engineer??research departments in Brazilian HEIs, the ties are teachers who teach?in more?than 1 research department.? paulo Em ter., 7 de jun. de 2022 ¨¤s 19:53, Steve Borgatti <steve.borgatti@...> escreveu:
--
Paulo Matui 12 98122 6600 |
¿ªÔÆÌåÓýYes, it uses the Tarjan depth-first search algorithmSteve Borgatti
LINKS Center for Social Network Analysis
On Jun 8, 2022, at 07:59, Paulo Matui <paulo.matui@...> wrote:
|
|
Professor, Great, which means that, if I multiply the result?over a subgraph?by (Nsub-1), I'll get the number of eulerian cycles where it is not isolated or strong by itself - the digraph case where the node has only in-degree or out-degree. So, it is feasible?to calculate the number of eulerian cycles that a node is part in its directional adjacent matrix,?including ties among the adjacent nodes. Then, the point is how to generate the adjacent subgraph? I can generate subgraph from a partition vector, but the best for my research is if there is a way to extract the adjacent submatrix of a node. The theoretical approach of this model comes from Padgett, J. F. (1997) - The emergence of simple ecologies of skills. The organizational learning process is treated as cycles. From the field side, some of the informants confirmed they use multiple allocation practices to induce results in quality, productivity?and relevance in the HEI research?community, some not. I want to triangulate objective?data, and perhaps correlates individual categories.? Em qua., 8 de jun. de 2022 ¨¤s 09:03, Steve Borgatti <sborgatti@...> escreveu:
--
Paulo Matui 12 98122 6600 |
Paulo Matui
