Papapetrou, Panagiotis and Chistiakova, T. and Hollmen, J. and Kalogeraki, V. and Gunopulos, D. (2012) Finding representative objects using link analysis ranking. In: UNSPECIFIED (ed.) Proceedings of the 5th International Conference on PErvasive Technologies Related to Assistive Environments. New York, USA: ACM Publications. ISBN 9781450313001.
Abstract
Link analysis ranking methods are widely used for summarizing the connectivity structure of large networks. We explore a weighted version of two common link analysis ranking algorithms, PageRank and HITS, and study their applicability to assistive environment data. Based on these methods, we propose a novel approach for identifying representative objects in large datasets, given their similarity matrix. The novelty of our approach is that it takes into account both the pair-wise similarities between the objects, as well as the origin and "evolution path" of these similarities within the dataset. The key step of our method is to define a complete graph, where each object is represented by a node and each edge in the graph is given a weight equal to the pairwise similarity value of the two adjacent nodes. Nodes with high ranking scores correspond to representative objects. Our experimental evaluation was performed on three data domains: american sign language, sensor data, and medical data.
Metadata
| Item Type: | Book Section |
|---|---|
| School: | Birkbeck Faculties and Schools > Faculty of Science > School of Computing and Mathematical Sciences |
| Depositing User: | Sarah Hall |
| Date Deposited: | 26 Jul 2013 11:09 |
| Last Modified: | 09 Aug 2023 12:34 |
| URI: | https://eprints.bbk.ac.uk/id/eprint/7851 |
Statistics
Additional statistics are available via IRStats2.
Archive Staff Only (login required)
