what-when-how
In Depth Tutorials and Information
separation degree. herefore, for information dissemination, a node that enjoys
a lower separation degree from a given node tends to have a good performance of
information dissemination.
Newman's study of scientific collaborations shows that if two authors have
more mutual co-authors before, then we can predict the probability of their col-
laborations is higher. For instance, we assume that two scientists have six previous
common collaborators, then the possibility of their collaboration is about three
times higher than the pair who only has two collaborators, and more than 200
times higher than the pair who has none. Newman deems that the probability of
two scientists' collaborations increases when the number of their previous common
collaborators rises.
In this way, Liben-Nowell and Keinberg use the neighbor metric to predict
future collaborations. P(a,b) is the probability of collaboration of two assigned sci-
entists a and b in the future. he formula is as follows:
P(a,b) =
|
N(a)
∩
N(b)|
Where N(a) and N(b) are the total number of neighbors of scientists a and b indi-
vidually. From this formula, the common set of neighbors of scientists a and b
depends on the network layout. By using neighbor metrics we can predict the prob-
ability of the two scientists' collaborations in the future.
12.6 ModelingSpreadofIdeasin
OnlineSocialNetworks[5]
IM means an instant message network, which is a type of dynamic online social
network. he major diference between the IM networks and traditional networks
is that we cannot presume that a random node is active. A node may be offline
(i.e., disconnected from all links with other neighbors) at a time when it is inactive.
For instance, in our real world, a person with a communicable disease who avoids
contact with others will prevent its spread.
12.6.1 Susceptibility
he susceptibility
δ
denotes the possibility of a node that can accept the content
of a message, thus becoming a message carrier. In reality, one decides whether or
not to accept content depending on its personal appeal or professional relevance.
Hence, people tend to accept a message if its contents have something in com-
mon with their own interests. A high susceptibility indicates that the message is
similar the person's other interests. A random value 1 and 0 can reflect a node's
susceptibility.
