# 《Six Degrees》的笔记-第58页

###### vieplivee (vvvvv)

读过 Six Degrees

- 页码：第58页 2011-01-27 13:28:31

刚才忙着去八卦Anatol Rapoport的生平，忘了记一个要紧的话题。Today when reading Six Degrees: The Science of a Connected Age, I came across an interesting notion that I think I'd share with you.Erdos-Reyni's theory of random graph, when applied to analyzing social network, is missing the key component of the latter, that is, the social network is not as random as assumed by Erdos and Reyni. Hereby Anatol Rapoport raised an interesting notion of "random-biased nets", which can be used to describe graphs that are basically random, but are biased by a handful of given principles. This reminds me of point-determining graphs, which are graphs in which no two vertices share the same adjacencies. Applied to social network, point-determining graphs can be used to describe such "random-biased nets" such that no two persons who are not acquainted with each other share the same set of friends. --- Even though the notion seems to be too werid to fit with reality, it seems worthwhile to explore. Also I believe such graphs can be applicable not only for social network analysis, but also for other graph-based relations, such as spread of disease. I'll keep my eyes open for finding such opportunities to apply my theoretical findings to interesting situations in real life. Besides, I just thought about this --- what if there is a fixed probability p such that a vertex V is adjacent to both W1 and W2 with probability p when W1 and W2 are adjacent to each other... Does this sound like an interesting topic to work on?

## 说明 · · · · · ·

表示其中内容是对原文的摘抄