Help with interpretation: Betweenness centrality score and infection status
The independent variable "betweenness centrality?score" in my data is ranged between 0 and 4.79 (on continuous scale) for undirected nodes, whereas the dependent/outcome variable "infection status present/absent" in my data is binary?i.e., 0 or 1. I performed elastic net regression to estimate the effect size and found a negative?coefficient value. Could someone from the group help me how to interpret this result?
Thank you
Shovon
|
Thanks a lot, Steve! This is really helpful. Can't imagine I forgot such a basic principle.?Appreciate your help!
Best, Simon
On Mon, Feb 6, 2023 at 5:35 PM Steve Borgatti < sborgatti@...> wrote:
Simon, if you had a categorical variable with two categories, such as male and female, and you created two dummies (‘ismale’, ‘isfemale’) and then used both as independent variables, the regression would fail
because of perfect multi-collinearity. You have to use all but one dummy variable. Could that be what happened here?
?
steve
?
Stephen P. Borgatti
Carol Martin Gatton Chair of Management
Gatton College of Business and Economics
University of Kentucky
?
?
?
Hi everyone,
I hope you are doing well. When I used UCINET to test a node-level hypothesis, something went wrong. My dependent variable is centrality, and my independent variables
are some dummy?variables and some continuous variables. When I include all dummy variables, all Beta, T, and c.Sig disappear, whereas all p.Sig are 1. When I reduce one dummy variable, it becomes normal. Could anyone explain how this happens? Thanks.
|
Simon, if you had a categorical variable with two categories, such as male and female, and you created two dummies (‘ismale’, ‘isfemale’) and then used both as independent variables, the regression would fail
because of perfect multi-collinearity. You have to use all but one dummy variable. Could that be what happened here?
?
steve
?
Stephen P. Borgatti
Carol Martin Gatton Chair of Management
Gatton College of Business and Economics
University of Kentucky
?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]> On Behalf Of
yx6wu via groups.io
Sent: Monday, February 6, 2023 15:20
To: [email protected]
Subject: [ucinet] UCINET Regression Problem #error #help
?
?
Hi everyone,
I hope you are doing well. When I used UCINET to test a node-level hypothesis, something went wrong. My dependent variable is centrality, and my independent variables
are some dummy?variables and some continuous variables. When I include all dummy variables, all Beta, T, and c.Sig disappear, whereas all p.Sig are 1. When I reduce one dummy variable, it becomes normal. Could anyone explain how this happens? Thanks.
|
Hi everyone,
?
I hope you are doing well. When I used UCINET to test a node-level hypothesis, something went wrong. My dependent variable is centrality, and my independent variables are some dummy?variables and some continuous variables. When I include all dummy variables, all Beta, T, and c.Sig disappear, whereas all p.Sig are 1. When I reduce one dummy variable, it becomes normal. Could anyone explain how this happens? Thanks.
?
Best,
Simon?
|
Re: Foreign Alphabets
#help
Is it a specific routine that is screwing up? If I enter data with these names into the dl editor using a recent version of UCINET, like this:
?
?
It produces this output:
?
IMPORT DATA VIA DL EDITOR
--------------------------------------------------------------------------------
?
?
?
??????????????? 1 2 3 4
????????????????B B S ?
????????????????? ? t d
????????????????r ? a e
????????????????d s r g
????????????????????l ?
????????????????????i r
????????????????????n d
????????????????????g??
????????????????- - - -
?????1???? B?rd 0 0 1 0
?????2???? B??s 1 0 0 0
?????3 Starling????????
?????4? ?deg?rd 1 1 0 0
?
4 rows, 4 columns, 1 levels.
?
Output UCINET file:????? ???????????????starling (C:\Users\Steve Borgatti\Dropbox\data\ucinet data\starling
?
?
--------------------------------------
Running time: 00:00:11 seconds.
Output generated: 02 Feb 23 10:54:42
?
And a further analysis looks like this:
?
->dsp degree(starling)
Normalized = no
See also: indegree and outdegree functions
?
?
?
???????????????
1 2 3
????????????????D O I
????????????????e u n
????????????????g t d
????????????????r d e
????????????????e e g
????????????????e g r
??????????????????r e
??????????????????e e
??????????????????e??
????????????????- - -
?????1???? B?rd 3 1 2
?????2???? B??s 2 1 1
?????3 Starling 1 0 1
?????4? ?deg?rd 2 2 0
?
4 rows, 3 columns, 1 levels.
?
?
So, if your version of ucinet is not from the last few years, then upgrade. If it is recent, tell me exactly what procedures you used to read in the data and/or analyze it. Perhaps older routines don’t understand
Unicode.
?
steve
?
Stephen P. Borgatti
Carol Martin Gatton Chair of Management
Gatton College of Business and Economics
University of Kentucky
?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]> On Behalf Of
starlingdavidhunter via groups.io
Sent: Thursday, February 2, 2023 04:02
To: [email protected]
Subject: [ucinet] Foreign Alphabets #help
?
?
Hello everyone!
Working with network datasets in Scandinavia presents a few issues concerning names. In short, characters such as umlauts or "o with stroke" aren't recognized.
For example, names like?B??s or?B?rd or ?deg?rd get quite mangled in the output. I either have to Anglicize the names (not always appreciated), use digraphs, e.g. use "oe" for "o with stroke", or manually replace them in the output.
Are there any solutions for this, specifically a way? to get UCINET to recognize these characters?
thoughtfully
Starling
Also working with a few datasets in the Middle East but given the above, I'll assume for now that dealing Arabic script is a complete non-starter!
|
Hello everyone!
Working with network datasets in Scandinavia presents a few issues concerning names. In short, characters such as umlauts or "o with stroke" aren't recognized.
For example, names like?B??s or?B?rd or ?deg?rd get quite mangled in the output. I either have to Anglicize the names (not always appreciated), use digraphs, e.g. use "oe" for "o with stroke", or manually replace them in the output.
Are there any solutions for this, specifically a way? to get UCINET to recognize these characters?
thoughtfully
Starling
Also working with a few datasets in the Middle East but given the above, I'll assume for now that dealing Arabic script is a complete non-starter!
|
Re: Exporting Network Members at Different Tie Levels
Whoops. So to try this with other data I tried the Southern Women reduced to 1-mode. Unpacking generates no output. When I looked up the Zacchary Karate Club (zackar) I see that it was composed of both a C and E matrix and unpacking pulls out the C matrix output above. So, obviously I need to convert the Southern Women as my test but am not sure how to proceed. Thanks again.
|
Re: Imputing missing data with xUCINET
Thank you Prof. Borgatti! The solution suggested by you works!
Chuding, you can make your own replacena. Assume your matrix is called X. Then these lines
?
Xt <- t(X)
X[(X)] = Xt[(X)]
?
will replace your missing rows with their corresponding columns. Note that if you have multiple missing rows, such as 1 and 2, you will still have a missing value for the tie between node 1 and node 2, since
neither one responded.
?
Note there are more sophisticated methods of imputing network data:
?
Krause, R. W., Huisman, M., Steglich, C., & Snijders, T. A. (2018, August). Missing network data a comparison of different imputation methods.
In?2018 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM)?(pp. 159-163). IEEE.
?
?nidar?i?, A., Ferligoj, A., & Doreian, P. (2012). Non-response in social networks: The impact of different non-response treatments on the stability
of blockmodels.?Social Networks,?34(4), 438-450.
?
Stephen P. Borgatti
Carol Martin Gatton Chair of Management
Gatton College of Business and Economics
University of Kentucky
?
?
?
Dear all,
I used to analyze social network data with UCINET. And recently I am trying to use the xUCINET package to do the analyses considering its flexibility. Now I have a question about imputing missing data with xUCINET. According to Borgatti, Everett, and Johnson
(2018):
“In the case of symmetric or undirected relations, a simple cure is to fill in any missing rows with the data found in the corresponding column. The assumption is that, if
the respondent had been able to answer, they would have listed all the actors that mentioned them. This may not be exactly right, but it will be more accurate than treating the missing values as zeros. UCINET has a command called REPLACENA within Matrix Algebra
to do this.”
However, in the current version of xUCINET, the xImputeMissingData() function only imputes data with density. I am wondering whether we can manually write some codes in xUCINET to realize the function of REPLACENA in UCINET? If yes, could you please give me
an example? Many thanks!
|
Hello
I'm having this message when I tried to do reciprocity. I received error messages for both NodeReciprocity and GroupReciprocity. The input dataset is an h file, and the method I used is "dyad-based. Any suggestion to fix this problem?? ?(Note: when I close the tab, I got the Reciprocity result). Thank you so much.?
Sincerely
Wah Wah Myint
|
Re: Imputing missing data with xUCINET
Chuding, you can make your own replacena. Assume your matrix is called X. Then these lines
?
Xt <- t(X)
X[is.na(X)] = Xt[is.na(X)]
?
will replace your missing rows with their corresponding columns. Note that if you have multiple missing rows, such as 1 and 2, you will still have a missing value for the tie between node 1 and node 2, since
neither one responded.
?
Note there are more sophisticated methods of imputing network data:
?
Krause, R. W., Huisman, M., Steglich, C., & Snijders, T. A. (2018, August). Missing network data a comparison of different imputation methods.
In?2018 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM)?(pp. 159-163). IEEE.
?
?nidar?i?, A., Ferligoj, A., & Doreian, P. (2012). Non-response in social networks: The impact of different non-response treatments on the stability
of blockmodels.?Social Networks,?34(4), 438-450.
?
Stephen P. Borgatti
Carol Martin Gatton Chair of Management
Gatton College of Business and Economics
University of Kentucky
?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]> On Behalf Of
CHUDING LING via groups.io
Sent: Wednesday, January 25, 2023 06:39
To: [email protected]
Subject: [ucinet] Imputing missing data with xUCINET
?
?
Dear all,
I used to analyze social network data with UCINET. And recently I am trying to use the xUCINET package to do the analyses considering its flexibility. Now I have a question about imputing missing data with xUCINET. According to Borgatti, Everett, and Johnson
(2018):
“In the case of symmetric or undirected relations, a simple cure is to fill in any missing rows with the data found in the corresponding column. The assumption is that, if
the respondent had been able to answer, they would have listed all the actors that mentioned them. This may not be exactly right, but it will be more accurate than treating the missing values as zeros. UCINET has a command called REPLACENA within Matrix Algebra
to do this.”
However, in the current version of xUCINET, the xImputeMissingData() function only imputes data with density. I am wondering whether we can manually write some codes in xUCINET to realize the function of REPLACENA in UCINET? If yes, could you please give me
an example? Many thanks!
|
Re: netdraw arrowheads not removed
Thanks so much, Steve—I was concerned that my issue was too trivial for any attention. I won’t be back at my PC in my office until sometime next week so I’ll wait for the official fix—and I may need some brief
instruction on what to do to upgrade (it’s been quite a while since I’ve messed with UCINet, thanks to the pandemic and not doing Zoom conferences).
?
Kate McCain
?
Hi Katherine, I took a look at the code and it turns out that the vna files don’t save the arrowhead info (nor look for it when reading). I’ve fixed that now, but won’t upload until the weekend so that I can
make some changes in ucinet at the same time. If you don’t want to wait, copy this exe:
?
?
and overwrite your existing netdraw.exe file (if you can find it!).
?
steve
?
Stephen P. Borgatti
Carol Martin Gatton Chair of Management
Gatton College of Business and Economics
University of Kentucky
?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]> On Behalf Of
McCain,Katherine via groups.io
Sent: Tuesday, January 24, 2023 16:14
To: [email protected]
Subject: [ucinet] netdraw arrowheads not removed
?
?
I've been having some odd interactions with Netdraw--both when I open a UCINet data file or the original dl file. The original files themselves are fine --UCINet likes my dl file. I can open both data formats in Netdraw and work with the
results--rearranging the individual networks for visibility, etc. When I first open a file, the links have arrowheads at both ends. No problem -- I turn them off.
However, if I save the Netdraw data for that configuration as a vna file, when I reopen it later on, I have arrowheads on the links that will not turn off. It's frustrating.
Any thoughts?? It's a PITA to have to open a new network from the original data and redo all my work every time I want to tweak something (e.g. color/shape of nodes, vary tie size visualization, you name it).
Thanks in advance.
Katherine W. McCain, Professor Emerita
Department of Information Science
College of Computing & Informatics
Drexel University
|
Re: netdraw arrowheads not removed
Hi Katherine, I took a look at the code and it turns out that the vna files don’t save the arrowhead info (nor look for it when reading). I’ve fixed that now, but won’t upload until the weekend so that I can
make some changes in ucinet at the same time. If you don’t want to wait, copy this exe:
?
?
and overwrite your existing netdraw.exe file (if you can find it!).
?
steve
?
Stephen P. Borgatti
Carol Martin Gatton Chair of Management
Gatton College of Business and Economics
University of Kentucky
?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]> On Behalf Of
McCain,Katherine via groups.io
Sent: Tuesday, January 24, 2023 16:14
To: [email protected]
Subject: [ucinet] netdraw arrowheads not removed
?
?
I've been having some odd interactions with Netdraw--both when I open a UCINet data file or the original dl file. The original files themselves are fine --UCINet likes my dl file. I can open both data formats in Netdraw and work with the
results--rearranging the individual networks for visibility, etc. When I first open a file, the links have arrowheads at both ends. No problem -- I turn them off.
However, if I save the Netdraw data for that configuration as a vna file, when I reopen it later on, I have arrowheads on the links that will not turn off. It's frustrating.
Any thoughts?? It's a PITA to have to open a new network from the original data and redo all my work every time I want to tweak something (e.g. color/shape of nodes, vary tie size visualization, you name it).
Thanks in advance.
Katherine W. McCain, Professor Emerita
Department of Information Science
College of Computing & Informatics
Drexel University
|
Imputing missing data with xUCINET
Dear all,
I used to analyze social network data with UCINET. And recently I am trying to use the xUCINET package to do the analyses considering its flexibility. Now I have a question about imputing missing data with xUCINET. According to Borgatti, Everett, and Johnson (2018):
“In the case of symmetric or undirected relations, a simple cure is to fill in any missing rows with the data found in the corresponding column. The assumption is that, if the respondent had been able to answer, they would have listed all the actors that mentioned them. This may not be exactly right, but it will be more accurate than treating the missing values as zeros. UCINET has a command called REPLACENA within Matrix Algebra to do this.”
However, in the current version of xUCINET, the xImputeMissingData() function only imputes data with density. I am wondering whether we can manually write some codes in xUCINET to realize the function of REPLACENA in UCINET? If yes, could you please give me an example? Many thanks!
|
netdraw arrowheads not removed
I've been having some odd interactions with Netdraw--both when I open a UCINet data file or the original dl file. The original files themselves are fine --UCINet likes my dl file. I can open both data formats in Netdraw and work with the results--rearranging the individual networks for visibility, etc. When I first open a file, the links have arrowheads at both ends. No problem -- I turn them off.
However, if I save the Netdraw data for that configuration as a vna file, when I reopen it later on, I have arrowheads on the links that will not turn off. It's frustrating.
Any thoughts?? It's a PITA to have to open a new network from the original data and redo all my work every time I want to tweak something (e.g. color/shape of nodes, vary tie size visualization, you name it).
Thanks in advance.
Katherine W. McCain, Professor Emerita
Department of Information Science
College of Computing & Informatics
Drexel University
|
Re: Exporting Network Members at Different Tie Levels
Good that you included screenshot. You need to remove the period. So you just type
?
unpack zackar ???
?
and hit enter
?
steve
?
Stephen P. Borgatti
Carol Martin Gatton Chair of Management
Gatton College of Business and Economics
University of Kentucky
?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]> On Behalf Of
Scott Thomas via groups.io
Sent: Saturday, January 21, 2023 12:13
To: [email protected]
Subject: Re: [ucinet] Exporting Network Members at Different Tie Levels
?
?
Looked like a good solution but am having problem executing. I get a an error (scratch file not found). Two pics to illustrate. Never tried CLI before.
 
|
Re: Exporting Network Members at Different Tie Levels
Looked like a good solution but am having problem executing. I get a an error (scratch file not found). Two pics to illustrate. Never tried CLI before.  
|
Re: First time using geographical coordinates in Ucinet
Hello
Could you send me a pic to show how? please!
Dra. Alma Beatriz Navarro Cerda Profesora-Investigadora Facultad de Economía y Relaciones Internacionales Universidad Autónoma de Baja California
El jueves, 18 de noviembre de 2021, 15:10:09 GMT-8, Scott Thomas <scott.thomas99@...> escribió:
I have found just using the lat and long with value (+/-) and decimal works fine for UCINET and many other programs.
|
Re: Exporting Network Members at Different Tie Levels
Well, this isn’t very elegant, but how about this? Using the std ucinet dataset “zackar” as an example, type these commands into the CLI:
?
unpack zackar ???//this just creates a valued dataset called zachc
?
d0 = dichot(zachc gt 0)
d1 = dichot(zachc gt 1)
d2 = dichot(zachc gt 2)
d3 = dichot(zachc gt 3)
d4 = dichot(zachc gt 4)
d5 = dichot(zachc gt 5)
d6 = dichot(zachc gt 6)
d7 = dichot(zachc gt 7)
d = joinmat(d0 d1 d2 d3 d4 d5 d6 d7)
deg = degree(d)
iso = dichot(deg eq 0)
dsp iso
?
the result will be:
?
?
????????? 1 2 3 4 5 6 7 8
??????????D D D D D D D D
??????????0 1 2 3 4 5 6 7
??????????- - - - - - - -
????????1 0 0 0 0 0 1 1 1
????????2 0 0 0 0 0 0 1 1
????????3 0 0 0 0 0 0 1 1
????????4 0 0 0 1 1 1 1 1
????????5 0 0 0 1 1 1 1 1
????????6 0 0 0 0 0 1 1 1
????????7 0 0 0 0 0 1 1 1
????????8 0 0 0 0 1 1 1 1
????????9 0 0 0 0 0 1 1 1
???????10 0 0 1 1 1 1 1 1
???????11 0 0 0 1 1 1 1 1
???????12 0 0 0 1 1 1 1 1
???????13 0 0 0 1 1 1 1 1
???????14 0 0 0 0 0 1 1 1
???????15 0 0 0 1 1 1 1 1
???????16 0 0 0 0 1 1 1 1
???????17 0 0 0 1 1 1 1 1
???????18 0 0 1 1 1 1 1 1
???????19 0 0 1 1 1 1 1 1
???????20 0 0 1 1 1 1 1 1
???????21 0 0 0 1 1 1 1 1
???????22 0 0 1 1 1 1 1 1
???????23 0 0 0 1 1 1 1 1
???????24 0 0 0 0 0 1 1 1
???????25 0 0 0 1 1 1 1 1
???????26 0 0 0 0 0 0 0 1
???????27 0 0 0 0 1 1 1 1
???????28 0 0 0 0 1 1 1 1
???????29 0 0 1 1 1 1 1 1
???????30 0 0 0 0 1 1 1 1
???????31 0 0 0 1 1 1 1 1
???????32 0 0 0 0 0 0 0 1
???????33 0 0 0 0 0 1 1 1
???????34 0 0 0 0 0 1 1 1
?
The ‘1’ in row 1 col 6 indicates that person 1 became an isolate when we dichotomized at tie strength > 5
?
steve
?
Stephen P. Borgatti
Gatton Endowed Chair of Management
Gatton College of Business and Economics
University of Kentucky
?
toggle quoted message
Show quoted text
From: [email protected] < [email protected]> On Behalf Of
Scott Thomas via groups.io
Sent: Monday, January 16, 2023 16:58
To: [email protected]
Subject: [ucinet] Exporting Network Members at Different Tie Levels
?
?
To study network structure I currently use Netdraw to study a network by slowing increasing the tie number to see who drops out, the number of total network ties, and who remains. Is there a way to generate an output file simply listing
all the nodes and the tie level they become isolates? Thank you.
|
Exporting Network Members at Different Tie Levels
To study network structure I currently use Netdraw to study a network by slowing increasing the tie number to see who drops out, the number of total network ties, and who remains. Is there a way to generate an output file simply listing all the nodes and the tie level they become isolates? Thank you.
|
Re: shocking problems about subgroups-louvain method !!!
#help
I am not a mathematician but it appears from this link below that Louvain is focused on measuring modularity from non-modular to fully modular by measuring the density of edges inside vs outside a community. The algorithm listed in the link definitely uses edge weights but perhaps in a way that would explain your lack of differences?
https://en.wikipedia.org/wiki/Louvain_method
You also might like to try VOS as it has some mathematical advantages but in my experience so far I have found no significant difference in results with Louvain vs VOS. Both approaches are available in Pajek and VOS.
http://mrvar.fdv.uni-lj.si/pajek/community/LouvainVOS.htm
http://mrvar.fdv.uni-lj.si/pajek/community/LouvainVOS1.htm
|
