Linear fuzzy cluster extraction from non-euclidean relational data
Katsuhiro Honda, Takeshi Yamamoto, Naoki Haga, Akira Notsu, Hidetomo Ichihashi · World Automation Congress · 2010
How to handle relational data is an active topic in fuzzy clustering. This paper proposes an extended version of linear fuzzy clustering based on Fuzzy c-Medoids (FCMdd), which is used with Non-Euclidean relational data. In order to estimate the clustering criterion of distances between objects and linear prototypes using mutual non-Euclidean distances, a modification used in NERF (non-Euclidean-type Fuzzy c-Means) is applied to the relational data before FCMdd-type linear cluster extraction. An experimental result demonstrates that we can find a suitable set of medoids, which are used for spanning prototypical lines, even when the relational measure is not Euclidean.