Fast Algorithm for Integrating Clustering with Ranking on Heterogeneous Graphs

Yamazaki, Kotaro; Sato, Tomoki; Shiokawa, Hiroaki; Kitagawa, Hiroyuki

doi:10.1145/3282373.3282376

Cited by 2 publications

(9 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• Effectiveness: For further improving for the clustering speed, we extended our algorithm to utilize thread-based par-allelization on a modern manycore processor. • Highly accurate: Although our proposed algorithms do not compute the entire graph, their clustering results are more accurate than those of the state-of-the-art algorithm [29]. Additionally, our proposals output are almost the same as the clustering results of the RankClus for real-world graphs.…”

Section: Our Approaches and Contributionsmentioning

confidence: 80%

“…This property allows users to deploy the graph clustering algorithm in real-world applications easily. Our experimental analysis shows that our proposed algorithms are 3.5 times faster than the competitive algorithms [26], [29], [30], while outputting almost the same clustering results as RankClus. The RankClus framework can enhance the qualities of various web-based applications, however it is difficult to apply to large-scale bipartite graphs due to its exhaustive computations.…”

Section: Our Approaches and Contributionsmentioning

confidence: 96%

“…• Efficient: Our proposals outperform the RankClus [26] and the state-of-the-art algorithm [29] in terms of running time for real-world bi-type information networks. • Effectiveness: For further improving for the clustering speed, we extended our algorithm to utilize thread-based par-allelization on a modern manycore processor.…”

Section: Our Approaches and Contributionsmentioning

confidence: 98%

“…Additionally, our proposals output are almost the same as the clustering results of the RankClus for real-world graphs. • Easy to deploy: Our proposed methods require fewer userspecified parameters than the state-of-the-art algorithm [29]. This property allows users to deploy the graph clustering algorithm in real-world applications easily.…”

Section: Our Approaches and Contributionsmentioning

confidence: 99%

“…Hence, RankClus requires exhaustive computational costs to find converged clusters in a large-scale bi-type information networks. Recently, Yamazaki et al, proposed edgepruning approaches [29], [30] to reduce the computational costs of the RankClus framework. However, the improvements of efficiency are limited and these approaches sacrifice clustering quality compared to the RankClus framework.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Fast and Parallel RankClus Algorithm based on Dynamic Rank Score Tracking

Yamazaki

Matsugu

Shiokawa

et al. 2020

Journal of Information Processing

Self Cite

View full text Add to dashboard Cite

Given a bi-type information network, which is an extended model of well-known bipartite graphs, how can clusters be efficiently found in graphs? Graph clustering is now a fundamental tool to understand overviews from graph-structured data. The RankClus framework accurately performs clustering for bi-type information networks using ranking-based graph clustering techniques. It integrates graph ranking algorithms such as PageRank or HITS into graph clustering procedures to improve the clustering quality. However, this integration incurs a high computational cost to handle large bi-type information networks since RankClus repeatedly computes the ranking algorithm for all nodes and edges until the clustering procedure converges. To overcome this runtime limitation, we present a novel RankClus algorithm that reduces the running time for large bi-type information networks. Our proposed method employs dynamic graph processing techniques into the iterative ranking procedures included in RankClus. By dynamically updating ranking results, our proposal reduces the number of computed nodes and edges during repeated ranking procedures. For further improving the efficiency, we also present a parallel implementation of our proposed algorithm by using thread-based parallelization on a modern manycore CPU. We experimentally verify using real-world datasets that our fast and parallel proposed methods successfully reduces the running time while maintaining the clustering quality of RankClus.

show abstract

Section: Our Approaches and Contributionsmentioning

confidence: 80%

Section: Our Approaches and Contributionsmentioning

confidence: 96%