Community search is a fundamental graph mining task. Unfortunately, most previous community studies focus mainly on identifying communities in network without temporal information. In this paper, we study the problem of finding persistent network, which every edge associated with timestamp. Our goal to identify that are over time. To end, propose novel model called (θ,τ) community. We prove maximum k-core NP-hard. solve problem, branch and bound algorithm several carefully-designed pruning...

Enumerating maximal cliques from an uncertain graph is a fundamental problem in analysis. Given G, set of nodes C G (k, τ)-clique if (1) |C|>k and clique with probability at least τ, (2) node meeting (1). The state-of-the-art algorithm for enumerating all τ)-cliques very costly when handling large graphs, as its time complexity proportional to 2^n where n the number graph. To overcome this issue, we propose two new core-based pruning algorithms reduce size without missing any τ)-clique. We...

A (p,q)-biclique is a complete subgraph (X,Y) that |X|=p, |Y|=q. Counting (p,q)-bicliques in bipartite graphs an important operator for many graph analysis applications. However, getting the count of large p and q (e.g., p,q ≥ 10) extremely difficult, because number increases exponentially with respect to q. The state-of-the-art algorithm this problem based on enumeration technique which often costly due exponential blowup space (p,q)-bicliques. To overcome problem, we first propose novel...

Mining cohesive subgraphs from a network is fundamental problem in analysis. Most existing subgraph models are mainly tailored to unsigned networks. In this paper, we study the of seeking signed network, which each edge can be positive or negative, denoting friendship conflict respectively. We propose novel model, called maximal (α, k)-clique, that represents Specifically, k)-clique clique every node has at most. negative neighbors and least [αk] (α ≥ 1). show enumerating all k)- cliques...

Maximal clique enumeration is a fundamental operator in graph analysis. The model of clique, however, typically too restrictive for real-world applications as it requires an edge every pair vertices. To remedy this restriction, practical analysis often resort to find relaxed cliques alternatives. In work, we investigate notable model, called s-defective which allows at most s edges be missing. Similar the complexity maximal enumeration, problem enumerating all also NP-hard. solve problem,...

Finding cohesive subgraphs from a bipartite graph is fundamental operator in analysis. In this paper, we focus on the problem of mining that satisfy hereditary property. Here subgraph meets property if all its same as itself. We show several important models, such maximal biclique and k-biplex, The enumerating was known to be NP-hard. To solve problem, first propose novel general pivot-based enumeration framework efficiently enumerate graph. Then, based our framework, develop new algorithm...

Personalized PageRank (PPR) computation is a fundamental problem in graph analysis. The state-of-the-art algorithms for PPR are based on bidirectional framework which include deterministic forward push and Monte Carlo sampling procedure. procedure, however, often has relatively-large variance, thus reducing the performance of algorithms. To overcome this issue, we develop two novel variance-reduced techniques computation. Our first technique to apply power iterations reduce variance We prove...

Computing the personalized PageRank vector is a fundamental problem in graph analysis. In this paper, we propose several novel algorithms to efficiently compute with decay factor α based on an interesting connection between values and weights of random spanning forests graph. Such derived newly-developed matrix forest theorem graphs. Based this, present efficient sampling algorithm via simulating loop-erased α-random walks estimate vector. Compared all existing methods, striking feature our...

Maximal clique enumeration on uncertain graphs is a fundamental problem in graph analysis. In this paper, we study of enumerating all maximal (k,n)-cliques an G, where vertex set H G (k,n)-clique if (1) (|H| ≥ k) with probability no less than n, and (2) satisfying (1). The state-of-the-art algorithms for are based technique which often very costly. This because the techniques may explore subsets (k,n)-clique, thus resulting many unnecessary computations. To overcome issue, propose several...

Resistance distance is a fundamental metric to measure the similarity between two nodes in graphs which has been widely used many real-world applications. In this paper, we study problems on approximately computing resistance distance: (i) single-pair query aims at calculating r(s, t) for given pair of (s, t); and (ii) single-source compute all distances u) u graph with source node s. Existing algorithms these are often costly large graphs. To efficiently solve problems, first establish...

Mining cohesive subgraphs from a network is fundamental problem in analysis. Most existing subgraph models are mainly tailored to unsigned networks. In this paper, we study the of seeking signed network, which each edge can be positive or negative, denoting friendship conflict, respectively. We propose novel model, called maximal (a, k)-clique, that represents Specifically, (α, k)-clique clique every node has at most k negative neighbors and least ⌈ak⌉ (α ≥ 1). show enumerating all...

A k-biplex is an induced subgraph of a bipartite graph which requires every vertex on the one side disconnecting at most k vertices other side. Enumerating all maximal k-biplexes in fundamental operator analysis and finds applications various domains, including community detection, online recommendation, fraud detection finance networks. The state-of-the-art solutions for enumeration suffer from efficiency issues as increases (k ≥ 2), with time complexity O(m 2 n ), where (m) denotes number...

Effective resistance (ER) is a fundamental metric for measuring node similarities in graph, and it finds applications various domains including graph clustering, recommendation systems, link prediction, neural networks. The state-of-the-art algorithm computing effective relies on landmark technique, which involves selecting that easy to reach by all the other nodes as landmark. performance of this technique heavily depends chosen node. However, many real-life graphs, not always possible find...

The study of k -defective cliques, defined as induced subgraphs that differ from cliques by at most missing edges, has attracted much attention in graph analysis due to their relevance various applications, including social network and implicit interaction predictions. However, determining the maximum clique graphs been proven be an NP-hard problem, presenting significant challenges finding efficient solution. To address this we develop a theoretically practically algorithm leverages...

Core decomposition on uncertain graphs is a fundamental problem in graph analysis. Given an G, the core to determine all (k, \eta)-cores where \eta)-core maximal subgraph of G such that each node has \eta-degree no less than k within subgraph. The state-of-the-art algorithm for solving this based peeling technique which iteratively removes nodes with smallest \eta-degrees and also dynamically updates their neighbors' \eta-degrees. Unfortunately, we find dynamical updating incorrect due...

Finding all maximal k-plexes on networks is a fundamental research problem in graph analysis due to many important applications, such as community detection, biological analysis, and so on. A k-plex subgraph which every vertex adjacent but at most k vertices within the subgraph. In this paper, we study of enumerating large develop several new efficient techniques solve problem. Specifically, first propose novel upper-bounding prune unnecessary computations during enumeration procedure. We...

K-clique counting is a fundamental problem in network analysis which has attracted much attention recent years. Computing the count of k-cliques graph for large k (e.g., = 8) often intractable as number increases exponentially w.r.t. (with respect to) k. Existing exact k-clique algorithms are hard to handle dense graphs, while sampling-based solutions either require huge samples or consume very high storage space achieve satisfactory accuracy. To overcome these limitations, we propose new...

<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$K$</tex-math></inline-formula> -clique counting is a fundamental problem in network analysis which has attracted much attention recent years. Computing the count of notation="LaTeX">$k$</tex-math></inline-formula> -cliques graph for large (e.g., notation="LaTeX">$k=8$</tex-math></inline-formula> ) often intractable as number increases exponentially w.r.t....

As an essential equipment for high-speed precision machining, the motorised spindle is usually widely used in CNC machine tools and machining centres. However, due to its compact structure, heat dissipation ability lacking, which makes internal temperature of rise produce thermal expansion. To further improve spindle, it necessary study characteristics coolant flow prediction algorithm based on simulation data. This paper mainly establishes network transfer model carries out flow-heat-solid...

Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI

A signed graph is a where each edge receives sign, positive or negative. The model has been used in many real applications, such as protein complex discovery and social network analysis. Finding cohesive subgraphs graphs fundamental problem. k-plex common for which every vertex adjacent to all but at most k vertices within the subgraph. In this paper, we propose of size-constrained antagonistic graph. proposed guarantees that resulting subgraph can be divided into two sub-k-plexes, both have...