汪小帆网络科学和社会学习.pptVIP

* * * * As we have known, regular lattices are clustered, but do not exhibit the small-world effect. On the other hand, random graphs show the small-world effect, but do not show clustering. Thus, it seems that for a sparse network, large L always corresponds to large C, and small L always corresponds to small C. Watts Strogatz built a small world model which serves as a transition from the completely regular network to the completely random graph and they also found that there does exist a network model which is a small world with high clustering. * The small world model can be generated as follows:…… Rewiring within this context means shifting one end of the connection to a new node chosen at random from the whole network, with the constraints that any two different nodes cannot have more than one connection between them, and no node can have a connection with itself. 其中规定，任意两个不同的节点之间至多只能有一条边并且每一个节点都不能有边与自身相连。 * They plot the average path length and clustering coefficient of the small-world model as a function of the rewiring probability for N=1000.K=5. Both are normalized to their values for the original regular lattice (p = 0). They found that for small values of p, the networks retained their properties of being highly clustered, but the average path lengths dropped dramatically and they called these networks ‘small world networks’. For example, for p = 0.01, (so that only 1% of the edges in the graph have been randomly changed), the clustering coefficient is over 95% of what it would be for a regular graph, but the characteristic path length is less than 20% of what it would be for a regular graph. Therefore, it is sufficient to make several random rewirings to decrease the average path length significantly. On the other hand, several rewired links cannot crucially change the local clustering property of the network.These networks are called small-world networks 取值都在区间内，从而可以在同一幅图中清楚地表示这两个值，并且也易于看出当增加时和的下降速度采用对数刻度的好处就是可以把我们最感兴趣的小值区间放大（当然，大值区间相对缩小）最后，由于对每一个