¿ªÔÆÌåÓýOriana, you have a 2-mode matrix of 3500 rows (farmers) and 80 columns (technologies). You are considering converting to a 1-mode matrix of 3500 by 3500 farmers, where x(k,j) indicates the extent to which farmersws k and? j use the same technologies. There is nothing wrong with this, it makes perfect sense. ? You also wondered whether your 1-mode matrix of 3500 x 3500 farmers would be too big for ucinet. I tested this question by constructing a matrix of 3500 x 3500 and running several analyses. They all ran fine, so my conclusion was that you can do the analysis that you want in ucinet. ? Now, I did express some concern about visualizing the dataset. If it is very dense, it could be a lot of netdraw to handle. However, few minutes ago I just drew a network of more than 4000 nodes with high density, and it worked fine. So that may not be a problem either. ? steve ? From: ucinet@... <ucinet@...>
Sent: Thursday, June 21, 2018 08:43 To: ucinet@... Subject: Re: RE: [UCINET] How many relations can Ucinet process? ? ? Steve, thanks very much for your answer ? Maybe it is easier if I divide the database by states.? ? But I would like to know your opinion about that I mentioned before, of converting this 2-mode matrix, in a 1-mode matrix, and evaluated how the farmers are related through the use of technology. I must admit that 2-mode matrix is confusing to me. Oriana Universidad de C¨®rdoba Spain En mi¨¦rcoles, 20 de junio de 2018 23:36:48 CEST, 'Steve Borgatti' steve.borgatti@... [ucinet] <ucinet@...> escribi¨®: ? ? ? It¡¯s hard to answer that question. I just created a valued 3500x3500 matrix with random values. Then I created an attribute dataset with 3500 rows and 1 column. ?Then I ran Network|Cohesion|ego-alter similarity (homophily), and it ran fine (1 second). Then I ran beta centrality and it ran in 19 seconds. But I didn¡¯t dare open the network in netdraw (since it¡¯s a valued network, it has more than 12 million ties). ? steve ? From:
ucinet@... <ucinet@...>
? ? Hi everyone, I am carry out a research in the field of innovation in the livestock system in Mexico. I would like to evaluate the relationship between farmers, depending of what kind of technology they use. ? The matrix is: In the rows: there are 1.600 farmers In the columns: 80 different technologies that the farmers use. ? In this case, (1) represent use of technology while, (0) represent not use of technology. ? I am thinking of converting this 2-mode matrix, in a 1-mode matrix, and them evaluated how the farmers are related through the use of technology.. I would like to evaluate the homophily as well. Is possible that this database increase in about 3.500 farmers, and I don¡¯t know if I can use UCINET, can this software process a database like this? I would to know how many relations can Ucinet process? or Have I to use Pajek? ? thanks in advance, Oriana, Universidad de C¨®rdoba, Spain. ?
|
