LJ graph structure
Jun. 12th, 2001 09:30 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
snydersLJ graph structure is a miniature model of the web. The Game (described by avva here) starts at a random point in LJ graph and traverses friend links till you reach yourself. The aim is of course to do it in a minimal number of steps.
Some useful statistics:
average (over 30 random samples) in/out degree of a vertex is 6.8-9.6, out of 30 randomly chosen users 4 had no friends
average (over 26 samples in a random walk) in/out degree is 34. Variance is very high so 30 trials are obviously not enough.
Conclusions:
1. 50% of LJ users have less than 5 friends. These are new users with only few friends, ppl who prefer to write privately or inactive users.
2. Random walk quickly falls into heavy clusters-communities (20-100 ppl).
3. Structure of community clusters and inter-community links is of interest.
4. There are about 180000 users, 530 from RU
5. Age peak is at 16
Walk Example (out degrees): 1-6-17-29-40-25-6-52-36-100-44-53-14-12-20-7-39-43-29-47-39-64-69-11-40-53
This place has updated stats and old LJ graph (28-Feb-2001) [friend pairs] info. Full stats page is here.
Some useful statistics:
average (over 30 random samples) in/out degree of a vertex is 6.8-9.6, out of 30 randomly chosen users 4 had no friends
average (over 26 samples in a random walk) in/out degree is 34. Variance is very high so 30 trials are obviously not enough.
Conclusions:
1. 50% of LJ users have less than 5 friends. These are new users with only few friends, ppl who prefer to write privately or inactive users.
2. Random walk quickly falls into heavy clusters-communities (20-100 ppl).
3. Structure of community clusters and inter-community links is of interest.
4. There are about 180000 users, 530 from RU
5. Age peak is at 16
Walk Example (out degrees): 1-6-17-29-40-25-6-52-36-100-44-53-14-12-20-7-39-43-29-47-39-64-69-11-40-53
This place has updated stats and old LJ graph (28-Feb-2001) [friend pairs] info. Full stats page is here.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
Request #6440
Date: 2001-06-13 11:18 pm (UTC)1. just study the graph: in/out degree, behavior of the ratios of in/out degrees (deviation from symmetric relation), diameter of the graph, properties of a random walk in such graph...
2. write a program to detect clusters-communities and study these communities in terms of: leaders, sizes, max-edge cut between different communities and "gateways" between communities.
3. Maybe run sort of pagerank on the graph to detect authorities, etc.. In order to do more clever things with pagerank "interest" info may be needed, but this is just an idea for the far future.
Clustering
Date: 2001-06-14 10:45 am (UTC)http://citeseer.nj.nec.com/deshpande00selective.html
http://www8.org/w8-papers/4a-search-mining/trawling/trawling.html