3/18/10
CIS 112 ‐ Networked Life
Midterm Solutions
Questions 2,3,4,7 were graded by Mickey Brautbar
Questions 1,5,6,8 were graded by Kareem Amin
Question 1)
a.
False : Preferential Attachment does not, for example, explain high clustering coefficients.
b.
False : There are six categories: SCC, IN, OUT, TENDRILS, TUBES, DISC.
c.
True
d.
False : Empirically, people tend to use geographic information more towards the beginning
of the chain.
e.
False : The number of possible friendships among your friends grows like K squared.
f.
True
g.
False : No, but it might have a high hub score.
h.
False : Christakis is a sociologist (and a physician). Fowler is a political scientist.
i.
False: In a social setting, clustering coefficients measure how often a friends share a
mutual friend. Alternatively, it measures how many triangles are in the network. This does
not necessitate that the vertices necessarily have anything “similar” about them.
j.
False : In Gladwell’s terminology, a “connector” has high degree.
Grading policy
:
i.
Each wrong answer [‐1 point].
Question 2)
a.
“www.wheresgeorge.com” is an on‐line bill tracking website. A participant with a one dollar
bill logs into the website service and records the bill's number and current location.
b.
The authors are interested in tracking human movements across the USA. The locations of the
dollar bills are considered as a proxy to the real movement patterns of their owners.
c.
The main finding is that over a short time frame (up to two weeks) the distance traveled by a
bill, between its current location and the next, follows a power‐law distribution with an
exponent of roughly 1.6.
d.
Prof. Kleinberg describes a theoretical model where people are located on a two dimensional
grid and a person gets an additional k random long distance acquaintances according to a
power‐law distribution with exponent alpha. Kleinberg's shows that efficient navigation
based on local knowledge is only possible when alpha equals 2.
We can draw a connection between the paper and Kleinberg’s model by interpreting the
dollar bill travel distance as a long‐distance connection between acquaintances. The findings
of the authors of “The Scaling Laws of Human Travel”, that human migration patterns follow a
power law with exponent 1.6, can then be interpreted as an instantiation of the Kleinberg's
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Documentmodel with parameter alpha of 1.6. Under this interpretation one should then discuss
whether 1.6 is close enough to 2 to get an efficient navigation. Any argument is fine as long it
is stated clearly.
This is the end of the preview.
Sign up
to
access the rest of the document.
 Winter '14
 Web search engine, Google search, PageRank, Grading policy

Click to edit the document details