An influence power-based clustering approach with PageRank-like model
| Autor: | Yu Zhao, Rongmin Gao, Yongpo Jia, Li Liu, Jun Zhong, Ming Liu, Xiwei Chen |
|---|---|
| Rok vydání: | 2016 |
| Předmět: |
Clustering high-dimensional data
DBSCAN Fuzzy clustering Computer science Correlation clustering Constrained clustering 02 engineering and technology 010501 environmental sciences computer.software_genre 01 natural sciences ComputingMethodologies_PATTERNRECOGNITION Data stream clustering CURE data clustering algorithm 0202 electrical engineering electronic engineering information engineering Canopy clustering algorithm FLAME clustering 020201 artificial intelligence & image processing Data mining Cluster analysis computer Software k-medians clustering 0105 earth and related environmental sciences |
| Zdroj: | Applied Soft Computing. 40:17-32 |
| ISSN: | 1568-4946 |
| DOI: | 10.1016/j.asoc.2015.10.050 |
| Popis: | Graphical abstractDisplay Omitted HighlightsInfluence power is introduced to measure points' relation in and between clusters.PageRank-like algorithm is adopted to estimate influence power evolving over time.Clustering is performed in a tree-growing fashion based on influence power.The proposed method outperforms on seven complex and non-isotropic datasets.Alzheimers disease and race walking recognition datasets validates the applicability. In this paper, we present a clustering method called clustering by sorting influence power, which incorporates the concept of influence power as measurement among points. In our method, clustering is performed in an efficient tree-growing fashion exploiting both the hypothetical influence powers of data points and the distances among data points. Since influence powers among data points evolve over time, we adopt a PageRank-like algorithm to calculate them iteratively to avoid the issue of improper initial exemplar preference. The experimental results show that our proposed method outperforms four well-known clustering methods across seven complex and non-isotropic datasets. Moreover, our simple clustering method can be easily applied to several practical clustering problems. We evaluate the effectiveness of our algorithm on two real-world datasets, i.e. an open dataset of Alzheimers disease protein-protein interaction network and a dataset for race walking recognition collected by ourselves, and we find our method outperforms other methods reported in the literature. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect