After three lectures introducing SNA (Social Network Analysis), I got an overall understanding on SNA in social networking. SNA is very important in engineering community. You will never know how closed the relationship is between the reality and SNA.
Social Network Analysis is the study of the pattern of interaction between actors. A social network is a group of collaborating individuals or entities that are related to each other, and in SNA, Social Network is formally defined as a set of social actors, or nodes, members that are connected by one or more types of relations.
I learned a lot of concepts from the SNA lectures, such as Centrality and Prestige, these words are very easy to understand for their literal meanings. But how can we relate them with the mathematical problems? It's really an interesting topic.
In SNA, Centrality can be studied with undirected or directed graphs (Directed graphs can also be described as digraphs.) while Prestige can only be studied with directed graphs. The most impressive description by Professor Rosanna Chan was that being an actor with prestige, it doesn't mean you have to give out a lot of choices, but only need to receive a great deal of choices, and then you can be called to have prestige. I can give an example to make everyone know better about this concept.
This is a sociogram showing the interactions of the students in the IEMS5720 blogosphere. As you can see, Su Jing and Helei Cui are both the prominent actors who are extensively involved in the relationships with other students, that means they have much more ties with others, if you want to know more people, the access to their blog is a good choice, no one would choose the isolated ones in the picture as you can access to nobody else except the blog writer himself. So, Su Jing and Helei Cui can both be called the actors with high centrality because they have most ties with others and centrality doesn't care the directions. They are all recognized by others as a major channel of relational information. However, if you want to know who is the actor with highest prestige, you can't only count the ties on the picture, but have to pay attention to the arrows. It is apparent that Su Jing is the person who have the highest prestige. I can prove to you through some other pictures.
For her one blog, she soon received 10 comments from our classmates while only one comment the other student received. And also you can see, even if she didn't reply to them, she still received a lot of comments, that means she has prestige. Maybe because her former blogs were attractive and soon other students got to know this and they all went to read her blogs.
And then let's see the relationships between some centralized or decentralized networks and the mathematical formulas.
If we assume CA(ni) is an actor centrality index, CA(n*) is the largest among CA(ni), the whole actors count g.
Then the general centralization index is
There are three types of centrality, Degree, Closeness, and Betweenness.
Degree centrality is the sum of all others actors who are directly connected to the actor in concern,
Closeness represents the mean of the geodesic distances between some particular node and all other nodes connected with in. The group closeness centralization measures the overall level of closeness in a network, and the group closeness can be expressed as:
Betweenness Centrality means the number of times a node connects pairs od other nodes, who otherwise would not be able to reach one another.
I don't mean to express clearly the difference among them. But I want to point out the common features of Centrality.
The left network is called star and the right one is a circle.
According to the given formulas, we can get the answers that CD=CC=CB=1 for the star and CD=CC=CB=0 for the circle.
What does it mean? As I have mentioned above, these two are all networks, the star is called the centralized network and the circle is called the decentralized network.
CD=CC=CB=0 means that all actors have exactly the same centrality index, which is obvious from the circle, that's why it's called decentralized network.
CD=CC=CB=1 means that one actor completely dominate or overshadow the other actors, which is exactly the central point in the star network. And so it's called centralized network.
You can know every network's centrality through computing with these formulas, it doesn't matter if the networks are too complicated, the formulas can help you.
SNA is so interesting and full of knowledge that I am now eager to learn more in the near coming lectures.









