Efficient simrank-based similarity join over large graphs

Abstract

Graphs have been widely used to model complex data in many real-world applications. Answering vertex join queries over large graphs is meaningful and interesting, which can benefit friend recommendation in social networks and link prediction, etc. In this paper, we adopt "SimRank" to evaluate the similarity of two vertices in a large graph because of its generality. Note that "SimRank" is purely structure dependent and it does not rely on the domain knowledge. Specifically, we define a SimRank-based join (SRJ) query to find all the vertex pairs satisfying the threshold in a data graph G. In order to reduce the search space, we propose an estimated shortest-path distance based upper bound for SimRank scores to prune unpromising vertex pairs. In the verification, we propose a novel index, called h-go cover, to efficiently compute the SimRank score of a single vertex pair. Given a graph G, we only materialize the SimRank scores of a small proportion of vertex pairs (called h-go covers), based on which, the SimRank score of any vertex pair can be computed easily. In order to handle large graphs, we extend our technique to the partition-based framework. Thorough theoretical analysis and extensive experiments over both real and synthetic datasets confirm the efficiency and effectiveness of our solution.

Efficient partial-pairs simrank search on large networks

“…The best-known work [20] provides an excellent exposition of a threshold-based version that retrieves {(a, b)} ∀a∈A,∀b∈B whose SimRank score s(a, b) is above a given threshold θ. However, [20] is not self-contained as it relies on the premise that the SimRank scores with node-pairs in an h-go cover 2 must be known in advance.…”

“…However, [20] is not self-contained as it relies on the premise that the SimRank scores with node-pairs in an h-go cover 2 must be known in advance. Precisely, for a given graph G, first, [20] adopts a heuristic (i.e., Algorithms 2 and 3 of [20]) to find a portion of node-pairs on the tensor graph G ⊗ G (known as an h-go cover set).…”

“…However, [20] is not self-contained as it relies on the premise that the SimRank scores with node-pairs in an h-go cover 2 must be known in advance. Precisely, for a given graph G, first, [20] adopts a heuristic (i.e., Algorithms 2 and 3 of [20]) to find a portion of node-pairs on the tensor graph G ⊗ G (known as an h-go cover set). Then, the SimRank scores with node-pairs in this h-go cover set have to be provided as an input parameter TB (in Algorithms 1 and 4 of [20]) and materialized, aiming to find the similarities of other nodepairs in A × B but outside the h-go cover set.…”

“…Precisely, for a given graph G, first, [20] adopts a heuristic (i.e., Algorithms 2 and 3 of [20]) to find a portion of node-pairs on the tensor graph G ⊗ G (known as an h-go cover set). Then, the SimRank scores with node-pairs in this h-go cover set have to be provided as an input parameter TB (in Algorithms 1 and 4 of [20]) and materialized, aiming to find the similarities of other nodepairs in A × B but outside the h-go cover set. However, it is hard to know in advance the SimRank scores of the hgo cover set (i.e., the input parameter TB of Algorithms 1 and 4).…”

“…However, it is hard to know in advance the SimRank scores of the hgo cover set (i.e., the input parameter TB of Algorithms 1 and 4). In addition, there is no rigorous complexity analysis in [20]. This highlights our need to focus on the efficiency of assessing partial-pairs SimRank in a self-contained manner.…”