离线瞬态社会网络(offline ephemeral social network,OffESN)是一种在特定时间通过参加特定事件临时组建的新型社会网络.随着移动智能终端的普及和短距离通信技术(如蓝牙、RFID技术等)的发展,该类型网络得到工业界和学术界越来越多的关...离线瞬态社会网络(offline ephemeral social network,OffESN)是一种在特定时间通过参加特定事件临时组建的新型社会网络.随着移动智能终端的普及和短距离通信技术(如蓝牙、RFID技术等)的发展,该类型网络得到工业界和学术界越来越多的关注.位置邻近(location proximity)关系是指用户在离线网络中的相遇关系.针对位置邻近关系的动态变化性及持续时间短等特征,主要研究离线瞬态社会网络中多用户邻近关系预测问题.首先,给出离线瞬态社会网络中的相关概念及问题定义;然后,设计多用户邻近关系预测总体框架,包括网络片段收集、叠加网络构建、网络过滤及极大紧密子图发现等步骤.由于多邻近关系的数量及每个邻近关系中用户的数量不能事先确定,基于分裂思想提出了一种极大紧密子图挖掘策略,以预测多用户位置邻近关系.该挖掘算法是以加权边介数为网络分裂依据,以聚集密度为分裂结束条件.在2个真实数据集上完成了实验,验证了所提出预测策略的可行性及效率.展开更多
现实生活中的网络，如生物蛋白网络、无线传感器网络等都存在着很多不确定性，如何准确、快速地发现其中有效的信息具有特别重要的意义。由于发现前 K 个最紧密子图具有较高的复杂性并且实现条件较高，本文根据实际背景研究了从不确定...现实生活中的网络，如生物蛋白网络、无线传感器网络等都存在着很多不确定性，如何准确、快速地发现其中有效的信息具有特别重要的意义。由于发现前 K 个最紧密子图具有较高的复杂性并且实现条件较高，本文根据实际背景研究了从不确定图中发现存在概率较高的前 K 个紧密子图问题，分析不确定图的连通性和紧密子图存在概率，提出了不确定相对 K 紧密子图发现算法。在算法中，首先计算不确定图的连通指数，确定不确定阈值，根据不确定阈值计算子图存在概率，最终得到 K 个相对紧密子图。最后，通过若干组实验，验证了此算法可以高效、准确地发现不确定图中的紧密子图，能够解决生活中出现的各种问题。展开更多
On basis of two definitions that 1. an induced subgraph by a vertex vi E G and its neighbors in G is defined a vertex adjacent closed subgraph denoted by Qi （=G[V（Nvi）]）, with the vertex vi called the hub; 2. A r...On basis of two definitions that 1. an induced subgraph by a vertex vi E G and its neighbors in G is defined a vertex adjacent closed subgraph denoted by Qi （=G[V（Nvi）]）, with the vertex vi called the hub; 2. A r（k,1）-1 vertices connected graph is called a （k,l）-Ramsey graph denoted by RG（k,l）,if and only if 1. RG（k,l） contains only cliques of degree k-1, and its complement contains only cliques of degree l-l; 2. the intersect Qi∩Qj of any two nonadjacent vertices vi and vj of RG（k,1） contains Kk.2, and the intersect Qi∩Qj of any two nonadjacent vertices vi and vj of its complement RG（l,k） contains KI.2. Two theorems that theoreml ： the biggest clique in G is contained in some Qi of G, and theorem2： r（k,l）= [ V（RG（k,I）） I ＋1 are put forward and proved in this paper. With those definitions and theorems as well as analysis of property of chords a method for quick inspection and building RG（k,I） is proposed. Accordingly, RG（10,3） and its complement are built, which are respectively the strongly 29-regular graph and the strongly 10-regular graph on orders 40. We have tested RG（10,3） and its complement RG（3,10）,and gotten r（3,10）=41.展开更多
Two definitions are given that Definitionl： an induced subgraph by a vertex vie G and its neighbors in G is defined as a vertex adjacent closed subgraph, and denoted by Qi （=G[V（Nvi）]）, with the vertex vi called ...Two definitions are given that Definitionl： an induced subgraph by a vertex vie G and its neighbors in G is defined as a vertex adjacent closed subgraph, and denoted by Qi （=G[V（Nvi）]）, with the vertex vi called the hub; and Definition2： A r（k,I）-I vertices graph is called the （k,l）-Ramsey graph, denoted by RG（k,1）, if RG（k,1） only contains cliques Kk.1 and the intersect QiNQj of any two nonadjacent vertices vi and vj of RG（k,I） contains only Kk-2. Meanwhile, the RG（k,l）＇s complement RG（I,k） contains only cliques Kl.l, and the intersect QiNQj of any two nonadjacent vertices vi and vj of RG（I,k） contains only Ki.2. On the basis of those definitions, two theorems are put forward and proved in this paper. They are Theoreml： the biggest clique in G is contained in some Qi of G, and Theorem 2： r（k,1） = [V（RG（k,I））I ＋ 1. With those definitions and theorems as well as analysis of chord property, a method for quick inspection and building RG（k,1） is proposed. Accordingly, RG（4,6） is built, it is a strongly 14-regular graph on order 35. We have tested RG（4,6） and its complement, as a result, they meet the defintion2, so we proclaim that r（4,6）=36.展开更多
文摘离线瞬态社会网络(offline ephemeral social network,OffESN)是一种在特定时间通过参加特定事件临时组建的新型社会网络.随着移动智能终端的普及和短距离通信技术(如蓝牙、RFID技术等)的发展,该类型网络得到工业界和学术界越来越多的关注.位置邻近(location proximity)关系是指用户在离线网络中的相遇关系.针对位置邻近关系的动态变化性及持续时间短等特征,主要研究离线瞬态社会网络中多用户邻近关系预测问题.首先,给出离线瞬态社会网络中的相关概念及问题定义;然后,设计多用户邻近关系预测总体框架,包括网络片段收集、叠加网络构建、网络过滤及极大紧密子图发现等步骤.由于多邻近关系的数量及每个邻近关系中用户的数量不能事先确定,基于分裂思想提出了一种极大紧密子图挖掘策略,以预测多用户位置邻近关系.该挖掘算法是以加权边介数为网络分裂依据,以聚集密度为分裂结束条件.在2个真实数据集上完成了实验,验证了所提出预测策略的可行性及效率.
文摘现实生活中的网络，如生物蛋白网络、无线传感器网络等都存在着很多不确定性，如何准确、快速地发现其中有效的信息具有特别重要的意义。由于发现前 K 个最紧密子图具有较高的复杂性并且实现条件较高，本文根据实际背景研究了从不确定图中发现存在概率较高的前 K 个紧密子图问题，分析不确定图的连通性和紧密子图存在概率，提出了不确定相对 K 紧密子图发现算法。在算法中，首先计算不确定图的连通指数，确定不确定阈值，根据不确定阈值计算子图存在概率，最终得到 K 个相对紧密子图。最后，通过若干组实验，验证了此算法可以高效、准确地发现不确定图中的紧密子图，能够解决生活中出现的各种问题。
文摘On basis of two definitions that 1. an induced subgraph by a vertex vi E G and its neighbors in G is defined a vertex adjacent closed subgraph denoted by Qi （=G[V（Nvi）]）, with the vertex vi called the hub; 2. A r（k,1）-1 vertices connected graph is called a （k,l）-Ramsey graph denoted by RG（k,l）,if and only if 1. RG（k,l） contains only cliques of degree k-1, and its complement contains only cliques of degree l-l; 2. the intersect Qi∩Qj of any two nonadjacent vertices vi and vj of RG（k,1） contains Kk.2, and the intersect Qi∩Qj of any two nonadjacent vertices vi and vj of its complement RG（l,k） contains KI.2. Two theorems that theoreml ： the biggest clique in G is contained in some Qi of G, and theorem2： r（k,l）= [ V（RG（k,I）） I ＋1 are put forward and proved in this paper. With those definitions and theorems as well as analysis of property of chords a method for quick inspection and building RG（k,I） is proposed. Accordingly, RG（10,3） and its complement are built, which are respectively the strongly 29-regular graph and the strongly 10-regular graph on orders 40. We have tested RG（10,3） and its complement RG（3,10）,and gotten r（3,10）=41.
文摘Two definitions are given that Definitionl： an induced subgraph by a vertex vie G and its neighbors in G is defined as a vertex adjacent closed subgraph, and denoted by Qi （=G[V（Nvi）]）, with the vertex vi called the hub; and Definition2： A r（k,I）-I vertices graph is called the （k,l）-Ramsey graph, denoted by RG（k,1）, if RG（k,1） only contains cliques Kk.1 and the intersect QiNQj of any two nonadjacent vertices vi and vj of RG（k,I） contains only Kk-2. Meanwhile, the RG（k,l）＇s complement RG（I,k） contains only cliques Kl.l, and the intersect QiNQj of any two nonadjacent vertices vi and vj of RG（I,k） contains only Ki.2. On the basis of those definitions, two theorems are put forward and proved in this paper. They are Theoreml： the biggest clique in G is contained in some Qi of G, and Theorem 2： r（k,1） = [V（RG（k,I））I ＋ 1. With those definitions and theorems as well as analysis of chord property, a method for quick inspection and building RG（k,1） is proposed. Accordingly, RG（4,6） is built, it is a strongly 14-regular graph on order 35. We have tested RG（4,6） and its complement, as a result, they meet the defintion2, so we proclaim that r（4,6）=36.
