Abstract | ||
---|---|---|
In this article, we address the problem of automatic constraint selection to improve the performance of constraint-based clustering algorithms. To this aim we propose a novel active learning algorithm that relies on a k-nearest neighbors graph and a new constraint utility function to generate queries to the human expert. This mechanism is paired with propagation and refinement processes that limit the number of constraint candidates and introduce a minimal diversity in the proposed constraints. Existing constraint selection heuristics are based on a random selection or on a min-max criterion and thus are either inefficient or more adapted to spherical clusters. Contrary to these approaches, our method is designed to be beneficial for all constraint-based clustering algorithms. Comparative experiments conducted on real datasets and with two distinct representative constraint-based clustering algorithms show that our approach significantly improves clustering quality while minimizing the number of human expert solicitations. |
Year | DOI | Venue |
---|---|---|
2012 | 10.1016/j.patcog.2011.10.016 | Pattern Recognition |
Keywords | Field | DocType |
random selection,active query selection,existing constraint selection heuristics,comparative experiment,human expert,constraint candidate,constraint-based clustering algorithm,proposed constraint,new constraint utility function,human expert solicitation,automatic constraint selection | Data mining,Artificial intelligence,Constraint learning,Cluster analysis,Constraint satisfaction,Correlation clustering,Pattern recognition,Constraint graph,Constraint satisfaction dual problem,Constrained clustering,Machine learning,Mathematics,Hybrid algorithm (constraint satisfaction) | Journal |
Volume | Issue | ISSN |
45 | 4 | 0031-3203 |
Citations | PageRank | References |
18 | 0.62 | 36 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Viet-Vu Vu | 1 | 52 | 4.75 |
Nicolas Labroche | 2 | 139 | 17.87 |
Bernadette Bouchon-meunier | 3 | 1033 | 173.38 |