Chapter 3 - Basic topological properties of the network (complex network study notes)
Degree nodes and the average degree of
- Of: \ (node degree k i refers to the number of edges directly connected to node i \)
- The degree: $ I $ node pointing to other nodes in the number of edges
- Penetration: other node points to the node number of sides $ I $
- Average degree: The average of all the network nodes
- \ (K_i \) : the degree of node i
- \ (<K> \) : the average of the network
If the network is weighted G, then the degree of the node can be defined as the weighted 出强度
and入强度
Sparse and dense network of
- Density of the network: a containing density network $ N $ nodes $ \ Rho $ defined network actually exists in the number of edges $ M $ maximum possible ratio of the number of edges, i.e., \ (\ rho = \ frac { M } {\ frac {1} { 2} N (N-1)} \) for the Internet, the above formula can be removed 1/2.
If and when N tends to infinity and is a constant density of the network, the network indicates the actual number of sides and $ N ^ 2 $ is the same order, it is said that the network is dense.
- Average degree: \ (<K> = \ 2M FRAC {{N}} \)
- Density: \ (\ Rho = \ FRAC {M} {\ FRAC. 1} {2} {N (. 1-N)} \)
- And the relationship between the average density of: \ (<K> = (. 1-N) \ Rho \ approx N \ Rho \)
Average path length and diameter
Average path length
- The shortest path: between two nodes in the network
边数最少
path is called shortest path - From $ d_ $: is defined as the node i, j is the number of edges in the shortest path.
- Average path length $ L $: defined as the average distance of any two nodes in the network \ (L = \ frac {1 } {\ frac {1} {2} N (N-1)} \ sum_ {i> = j} d_ {ij} \)
Network diameter
- Network diameter D: the network is defined as the maximum distance of any two nodes \ (D = max (d_ { ij}) \)
In fact, we may be more concerned about is the distance between the vast majority of network users, so the first gives the following definition:
- \ (f (d) \) : network statistics from
等于
$ d $ of连通的节点对
the total network in连通的节点对
proportion - \ (G (d \) ): network statistics from
不超过
$ d $ of连通的节点对
the total network in连通的节点对
proportion
Generally, if the diameter $ D $ satisfies \ (g (D-1) <0.9, g (D) \ ge0.9 \) then said effective diameter D for the network.
Shortest path algorithm
- Dijkstra's algorithm: A weighted directed generally used network (nonnegative weight) of the shortest path between the nodes
- bellman-ford algorithm: for the presence of a negative weight value
Clustering coefficient (clustering coefficient)
- A node
聚类系数
depicts the node邻居节点
in任意一对节点
with even edges概率
. \ (C_i = clustering coefficient of a point = \ {the number of edges between neighboring nodes actually present point} frac {these neighbor nodes may exist the maximum number of edges} \) \ (C_i = \ {FRAC E_i} {K_i (-K_i. 1) / 2} = \ FRAC {} {2E_i K_i (-K_i. 1)} \)
among them
- \ (E_i \) : the number of edges between neighboring nodes actually present point
- $ K_i (k_i-1) / 2 $: the maximum number of edges which may exist neighbors
Distribution (degree distribution)
There will have a network connection, we are naturally concerned about the distribution of nodes in the network degrees.
Gaussian distribution (distribution is too / bell-shaped distribution)
Distribution is too positive for a continuous random variable, its corresponding discrete random variable, the most common is the Poisson distribution (Poisson Distribution) \ (P (K) = \ {FRAC \ ^ KE the lambda ^ {- \} the lambda } {k!} \)
Power law distribution (long-tailed distribution / unscaled distribution)
Power-law distribution and inspection, nature
Then access to information when available.