An somewhat erroneous statement I made in today's class..

In today's class, I said that it is possible to make poisson distribution as "peaked" as you want by
reducing the variance parameter (as part of making the point that the "drop off" in the power law distribution
is not as important as the "slow decay/long tail).

Well--not exactly. Poisson distribution, as you saw, has only *one* parameter--lambda, and the
means as well as the variance of the distribution will be lambda (this is unlike normal distribution,
for example, where you can set the mean and variance parameters indepdendently). The only way to reduce
the variance is thus to reduce the mean...  (you may also recall that for random networks,  lambda is p*N (i.e., the expected degree))

Here is a link on poisson distribution as well as how it is the limiting case of binomial distribution:

http://en.wikipedia.org/wiki/Poisson_distribution

cheers
rao

Speaking of the Long Tail...

http://www.thelongtail.com/the_long_tail/2005/09/long_tail_101.html

This article a while back sparked quite the frenzy in the business world...it can be seen as the direct cause of the renewed interest in non-Google engines and other businesses trying to capitalize on the Long Tail.

Project Deliverables

Just a reminder to bring a printout of your commented source code, example query results, and report to today’s class. 

 

We are not using digital drop box or any other electronic submission for the course…remaining homeworks/projects will also be submitted in class.

 

 

Garrett

 

 

Project Help - Sunday

It seems as though several people are interested in meeting today regarding the project so I will plan on being in the brickyard from around 3pm – 5pm.  I’ll plan on sitting in the second floor lab during this time (assuming I have access…otherwise I’ll be on the 5^th floor in which case you can just send me an email and I’ll come down and meet you on the second floor).

 

Garrett

 

 

CSE494/598 Project

After responding to several questions regarding the project, it seems there may be people with additional questions.  With the project deadline soon approaching, I wanted to see if anyone would find it useful if I were to make some time available tomorrow (Sunday) for you to come and ask questions face to face.    

 

For those who are seriously interested, please let me know.  If there are people expressing interest, then I can be available later in the afternoon (after 3pm or 4pm) Sunday for a couple hours. 

 

 

Thanks,

Garrett

 

 

Information Retrieve this!

SEO is a burgeoning business for those interested in seeing IR concepts taken to a business level...it also may be destroying the heart and soul of the Internet by turning a medium that's supposed to be about the free flow of information into a purely commercial endeavor, but that's probably just my own humble opinion. Anyway, check out some of the cool new stuff Google's doing that has that industry all atwitter! \

Thinking cap questions: Social networks..

Here are a couple of questions you can discuss (feel free to add other questions you want discussed too). You might want to look at the
blog to see other comments already made before you add your own. Also feel free to only answer the hard ones..


A. We considered three types of measures of centrality and prominence: based on degree, based on "closeness" to other entities in the network and
based on "between-ness"--i.e., how often is the entity going to be on the geo-desic path between other actor pairs. W.r.t. these, think of the following:

A0. Consider three canonical networks of size n: a star network, where one entity has (undirected) edges to all other entities; a circle network; and a line network.
 Think of what are the most "important" nodes in each of the networks, if we measure importance by degree based, closeness-based and between-ness based
notions respectively.

A1. are all these measures equally applicable to both directed and undirected social networks?

A2. In the case of directed social networks, can you attach different notions of importance to in-degree and out-degree?

A3. Can you think of generalizing the  in- and out- degree based notions to transitively take into account the importance of the actor from/to
which the edge is directed? (basically, you somehow want to say that the importance of a link from a node is proportional in some way to the
importance of the node itself).

A4. In the questions above, we kept silent about the label of the edge. Suppose in one network, the edges signify the relation "respects" and in
other network they signify the relation "hates". Does this difference throw a wrench into the kind of analysis the measures in 3 are trying to compute?

A5. Can you think of cases where closeness and between-ness measures of importance could be useful in the web-based networks (social networks, blogs or
page networks)?


B. We talked about small-world networks where the largest geo-desic path between any pair of
connected nodes in the network, also called the diameter of the network, is log of the size of the network.

B0: which of the networks in A0 are small-world networks by this definition?

B1: Suppose you are an actor in a small-world network of size N, and you only have local information (i.e., have access to only the links
incident on you). Suppose you are trying to look for a path to another actor/entity in the network with a certain "goal" property g().
We are interested in characterizing how much work you will need to do, in the worst case and the best case, to find this actor. Answer the
following:

B.1.1  What is the difference between this scenario and the usual graph-search problems you learned about in 310 or 471?

B.1.2. What is the best and worst case complexities (you decide how to measure complexity) of the task?

B.1.3. How does your answer to B.1.2. change if you have access to the whole network (i.e., you can see all the connections between all the nodes)


========that is all for now================