Editing Small World Networks 2008 (section)

=== Week 3&4 ===

Wenrui(Linda)

For the enron email date set, i understand that the enron.net file is where to store every single code and their email address. enron.pairs file is where to store who actually send to the email to whom, which presented by code, it shows the direction as well. md5-authors.txt gives every single email address a uniq code, with the same length and the same type. However, when i try to match them, i can't find some of the email address in the first file which is enron.net. is there any data missing? 

The enron.net Pajek file lists the nodes (number ID of person plus their email address) ''and'' a set of directed edges representing emails being sent. The other files are those I generated enron.net from. The message IDs are different to the person IDs (my software generates these). --[[User:Mattjb|Mattjb]] 14:01, 8 September 2008 (CST)

should i use java or C++ to make them look like what we need for the pajek software? just like what you had in md5-recipients.txt file? that will looks like Directed graph's file.

It's already in a Pajek format. I think there is still an issue with quote marks needed around the email addresses, and I will fix this up tonight. I can't test from home though. --[[User:Mattjb|Mattjb]] 14:01, 8 September 2008 (CST)

I have been trying to use the data from my facebook to make the file as well, i made the data set myself, what i did is make my name as the first node, all my friends in the facebook are the second level of the hierchical system, and then my friends' friends, and so on. it seems hard to get similar style of the file. facebook is where to know a friend, if you are a friend of some one, then the relationship should be both sides, so should be a undirected graph. 

Yes, it's undirected, because each time a person adds someone, that someone must confirm the friendship link. --[[User:Mattjb|Mattjb]] 14:01, 8 September 2008 (CST) 

it's obviously a hierachical. 

While it has levels, it's not hierarchical in the sense that any person has a position above another person, nor is it a tree structure, since you can have cycles (A is friends with B is friends with C is friends with A, etc.) --[[User:Mattjb|Mattjb]] 14:01, 8 September 2008 (CST) 

To anlaysis the third level of the data set, you need the data from the forth level, if i want to get a general result of the friendships, then that will be hard, because i will be need a really large data set, or maybe a data base. what's more, if just consider about the direct relationship but not 2 degrees or 3 degrees, in the hierachical system, we will only consider about the upper level and the lower level, then i don't know what the point to do that. 
However, if you find someone you know in your friend's list, add him/her to be your friend, that person may not know you at all. so in this situation the relationship is only one side.

For the matlab part, what kind of result should we show? i mean should we use matlab to draw the graphs to show all our value which we got from the pajek calculation?

Yes :) --[[User:Mattjb|Mattjb]] 14:01, 8 September 2008 (CST)

Zhujian(Sweet)

I think we got lots of social network. Should we do something about the Engineering network?

I have put some Internet networks up that you can manually turn into Pajek format files and process. Feel free to search for any other engineered networks. --[[User:Mattjb|Mattjb]] 14:01, 8 September 2008 (CST)

About my research, my friend and I just finished all the data collection. We are doing the analysis of these stuff. That is also about the social network. In my research, we just sort the links to 2 type, the "Money friendship" and "Emotion friendship". We will do the research of the 2 kind of links. I will do the analysis of data, my friend will do the part of literal analysis.
I think that is better than we just use the data from internet.

For the Engineering network, I prefer the telecommunication network, like the telephone links, the base stations distribution of mobile phone, the exchange of a city or province, atc. That is just my viewpoint. I think Linda can do the part of engineering. Choose some networks have the hierarchical directory structure and include other normal or special properties, then start ur work :)