Name
Abstract | ||
|---|---|---|
Clustering ensemble aims to obtain final clustering combining multiple diverse clustering solutions. It has already been established as an effective tool to yield a robust, accurate and stable consensus from the input clustering solutions. So far, a spectrum of approaches has already been proposed over the years to generate final ensemble from multiple solutions. One major drawback of most of the existing cluster ensemble approaches is that they require the final number of clusters as an input. In this paper, we propose a multi-objective optimization based algorithm for cluster ensemble problem that optimizes two objective functions simultaneously. The first objective is to maximize the overall similarity of the reference clustering solution to the input solutions, whereas the second objective is to minimize the standard deviation of the similarity values to avoid any bias. Moreover, in this proposed model, there is no need to supply the number of clusters a apriori to apply the algorithm, which is missing in most of the state-of-the-art approaches. The effectiveness of the proposed technique over the existing approaches is demonstrated by applying it on eight real-life datasets. |
Year | DOI | Venue |
|---|---|---|
2018 | 10.1109/SITIS.2018.00067 | SITIS |
Keywords | Field | DocType |
Clustering algorithms,Partitioning algorithms,Optimization,Measurement,Linear programming,Prediction algorithms,Bonding | Drawback,Data mining,Cluster (physics),Pattern recognition,Computer science,A priori and a posteriori,Prediction algorithms,Linear programming,Artificial intelligence,Cluster analysis,Standard deviation | Conference |
ISBN | Citations | PageRank |
978-1-5386-9385-8 | 0 | 0.34 |
References | Authors | |
0 | 3 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Sujoy Chatterjee | 1 | 0 | 0.34 |
| Nicolas Pasquier | 2 | 1840 | 140.08 |
| Anirban Mukhopadhyay | 3 | 711 | 50.07 |