基于综合游走策略的边嵌入链路预测算法

针对网络分析方法中研究的图数据默认使用节点图,只能得到节点的向量表示,不能直接将边表示成向量的问题,设计了一种基于有偏+无偏的图嵌入算法Line2Vec,并在此基础上提出基于边嵌入的链路预测框架(Line2Vec-L).首先,基于综合游走策略重新定义采样域节点的采样概率,并结合Word2Vec模型得到信息未被稀释、表示性强的节点图的边嵌入向量;然后,结合关联矩阵得到不存在边或未知边的向量表示,并将得到的边向量用于链路预测.实验结果表明Line2Vec在边向量表示上的有效性,并验证了Line2Vec-L的AUC值更高,由此说明采用Line2Vec可得到表示性更强的边向量,有助于提升链路预测的性能.

SOME RESULTSON R_2-EDGE-CONNECTIVITY OF EVEN REGULAR GRAPHS

Let Gbe a connected k(≥3)-regulargraph w ith girth g. A setSofthe edgesin G is called an R2-edge-cutifG- Sis disconnected and contains neither an isolated vertex nor a one- degree vertex. The R2-edge-connectivity of G, denoted by λ″(G), is the m inim um cardinality over allR2-edge-cuts, w hich is an im portantm easure for fault-tolerance of com puter intercon- nection netw orks. In this paper, λ″(G)= g(2k- 2) for any 2k-regular connected graph G(≠ K5) that is either edge-transitive or vertex-transitive and g≥5 is given.

On the Decay Number of a Graph

The decay number of a connected graph is defined to be the minimum number of the components of the cotree of the graph. Upper bounds of the decay numbers of graphs are obtained according to their edge connectivities. All the bounds in this paper are tight.Moreover, for each integer k between one and the upper bound, there are infinitely many graphs with the decay number k.

The Edge Connectivities of Total P_(k)-graphs

The P_(k)-path graph P_(k)(G)corresponding to a graph G has for vertices the set of all paths of length k in G.Two vertices are joined by an edge if and only if the intersection of the corresponding paths forms a path of length k-1 in G,and their union forms either a cycle or a path of length k+1.Let Ek={(v,p),p E V(P_(k)(G)),v is an end vertex of p in G},we define total P_(k)-graphs T_(k)(G)as Yk(G)=(V(G)UV(P_(k)(G)),E(G)U E(PI(G))U Ek).In this note,we introduce total P,-graphs Th(G)and study their edge connectivity,as the generaliza-tion of total graphs.

On the Adjacent Strong Edge Coloring of Halin Graphs

A proper k-edge coloring f of graph G(V, E) is said to be a k:-adjacent strong edge coloring of graph G(V,E) iff every uv∈E(G) satisfy f[u]≠f/[v], where f[u] = {f(uw)|uw ∈E(G)} then f is called k-adjacent strong edge coloring of G, is abbreviated k-ASEC: and x'as(G) = min{k|k-ASEC of G} is called the adjacent strong edge chromatic number. In this paper, we study the x'as(G) of Halin graphs with △A(G)≥5.

3-Restricted Edge Connectivity of Vertex Transitive Graphs of Girth Three

Let G be a k-regular connected graph of order at least six. If G has girth three, its 3-restricted edge connectivity λ3(G) ≤3k-6. The equality holds when G is a cubic or 4-regular connected vertex-transitive graph with the only exception that G is a 4-regular graph with λ3(G) = 4. Furthermore, λ3(G) = 4 if and only if G contains K4 as its subgraph.

Normal edge-transitive Cayley graphs on non-abelian groups of order 4p,where p is a prime number

We determine all connected normal edge-transitive Cayley graphs on non-abelian groups with order 4p, where p is a prime number. As a consequence we prove if IGI = 25p, δ = 0, 1, 2 and p prime, then F 1 Cay（G, S） is a connected normal 1/2 arc-transitive Cayley graph only if G = F4p, where S is an inverse closed generating subset of G which does not contain the identity element of G and F4p is a group with presentation F4p = （a, b ｜aP = b4 = 1, b-lab = a^λ）, where λ2 = -1 （mod p）.

On restricted edge-connectivity of replacement product graphs

This paper considers the edge-connectivity and the restricted edge-connectivity of replacement product graphs, gives some bounds on edge-connectivity and restricted edge-connectivity of replacement product graphs and determines the exact values for some special graphs. In particular, the authors further confirm that under certain conditions, the replacement product of two Cayley graphs is also a Cayley graph, and give a necessary and sufficient condition for such Cayley graphs to have maximum restricted edge-connectivity. Based on these results, we construct a Cayley graph with degree d whose restricted edge-connectivity is equal to d + s for given odd integer d and integer s with d 5 and 1 s d- 3, which answers a problem proposed ten years ago.

REFINED CONNECTIVITY PROPERTIES OF ABELIAN CAYLEY GRAPHS

Restricted edge connectivity of a graph G is defined to be the minimum size |U| of a set U of edges such that G-U is disconnected and G-U contains no trivial component K1. The high order edge connectivity Ni, i1, is the number of edge outsets of size i. TO determine all Ni, i 1, for a general graph is NP-hard. In this paper, the authors evaluated the restricted edge connectivity and the high order edge connectivity Ni, 1 i -1, for any connected Abelian Cayley graphs explicitly.

Super s-restricted edge-connectivity of vertex-transitive graphs

Let G be a connected graph with vertex-set V(G)and edge-set E(G).A subset F of E(G)is an s-restricted edge-cut of G if G-F is disconnected and every component of G-F has at least s vertices.Letλs(G)be the minimum size of all s-restricted edge-cuts of G andξs(G)=min{|[X,V(G)\X]|:|X|=s,G[X]is connected},where[X,V(G)\X]is the set of edges with exactly one end in X.A graph G with an s-restricted edge-cut is called super s-restricted edge-connected,in short super-λs,ifλs(G)=ξs(G)and every minimum s-restricted edge-cut of G isolates one component G[X]with|X|=s.It is proved in this paper that a connected vertex-transitive graph G with degree k>5 and girth g>5 is super-λs for any positive integer s with s 2g or s 10 if k=g=6.

TESTING k-EDGE-CONNECTIVITY OF DIGRAPHS

This paper presents an algorithm that tests whether a given degree-bounded digraph is k-edge-connected or E-far from k-edge-connectivity. This is the first testing algorithm for k-edge- connectivity of digraphs whose running time is independent of the number of vertices and edges. A digraph of n vertices with degree bound d is ε-far from k-edge-connectivity if at least εdn edges have to be added or deleted to make the digraph k-edge-connected, preserving the degree bound. Given a constant error parameter ε and a degree bound d, our algorithm always accepts all k-edge-connected digraphs and reiects all digraphs that is ε-far from k-edge-connectivity with orobabilitv at least 2/3.It runs in O（d（εd^-c）^k logεd^-1O）（c〉1 is a constant）time when input digraphs are restricted to be （k-1）-edge connected and runs in O（d（εd^-ck）^klogεd^-kO）（c〉1 is a constant）time for general digraphs.

Continuous-Time Independent Edge-Markovian Random Graph Process

In this paper, the continuous-time independent edge-Markovian random graph process model is constructed. The authors also define the interval isolated nodes of the random graph process, study the distribution sequence of the number of isolated nodes and the probability of having no isolated nodes when the initial distribution of the random graph process is stationary distribution, derive the lower limit of the probability in which two arbitrary nodes are connected and the random graph is also connected, and prove that the random graph is almost everywhere connected when the number of nodes is sufficiently large.