Skyline Community Search Algorithm Based On Triangle Connected K-truss

The community search was used to obtain compact subgraphs containing the query vertices from the data graph.This problem has been widely used in sociology,biology and other fields,and has received wide attention from researchers in recent years.In practice,the vertices in a graph may have multiple properties that need to be modeled using the property graph.Community search methods on existing property graphs return all non-dominant skyline communities and do not consider a community search for a specific query point.When dealing with community search problems based on query points using existing property graph,there are inconsistent results.This paper studies the skyline community search problem based on query points on the property graph and proposes the skyline community search model based on triangular connected k-truss.On this basis,two efficient online search algorithms and an index-based search algorithm are proposed,and the performance of the proposed algorithm is also experimentally verified.The specific research contents are as follows:To solve this problem,a new triangle-connected k-truss equivalent skyline community search model based on query points is proposed in this paper,including an edge-based BFS optimization algorithm and an algorithm based on TET-index index.Experiments are carried out based on real data sets,and the experimental results prove the effectiveness and correctness of the proposed algorithm.The specific contributions of this paper are as follows:First,TCT-I is a skyline community online search algorithm with triangular connected k-truss as the community structure on the attribute graph.The algorithm is mainly divided into two phases,the first one computing all triangular connected k-truss subgraphs containing the query vertex.The second phase is to calculate the skyline community based on the triangular connected k-truss subgraph of the first step.The basic idea is to get the skyline community results by capturing the communities with the highest impact in each dimension.Specifically,the non-decreasing ranking according to the attribute values of the vertices in the current dimension sub-graph can find the community that satisfies the triangular connected k-truss and has the greatest influence.Then we use the idea of the partition to update the candidate set,reduce the search space,and accelerate the computation of the next dimension.Further,to reduce the time consumption of the computational process,an edge-deletion optimization algorithm based on TCT-BD is proposed,which avoids the repeated calculation in the process of finding the maximum attribute value of each dimension.Second,we propose a community search algorithm based on the TCT index.In the first step,for each k value,all the triangle-connected trans subgraphs in the graph are calculated,and each trans subgraph is stored as a supemode.In the second step,use the tree index structure to maintain the inclusion relationship between the corresponding vertex sets,so that you can remove redundant storage and reduce the index size.The third step is to establish the mappingFinally,based on twelve real data sets,the experimental results verify the effectiveness and efficiency of the proposed algorithm.