An Extended Version of the Gustafson‐Kessel Algorithm for Evolving Data Stream Clustering

Autonomous data stream clustering implementing split-and-merge concepts – Towards a plug-and-play approach

“…Several incremental learning methods for prototype-based clustering (mostly in single-pass manner) have been proposed in recent literature, for instance single-pass k-means [33], dynamic fuzzy k-nearest neighbors clustering [44], a recursive variant of subtractive clustering termed as eClustering [3], a recursive Gustafson-Kessel approach [29] and an evolving version of it [34], evolving neural-type models based on neural gas [88], evolving self-organizing maps (ESOM) [26], or the approach in [92] within the application of time series data, to name a few -for a recent overview and comprehensive list of references, see [15,36]. Most of these extract ellipsoidal clusters in main position (axes-parallel).…”

Dynamic Data-Driven Fuzzy Modeling of Software Reliability Growth

“…Let us underline that the evolving clustering algorithm works with the currently available data x k . The introduced evolving clustering algorithm [7] assumes that the boundary of each cluster is defined by a cluster radius. The radius r i of the i-th cluster is equal to the maximal distance between the cluster centre v i and the points belonging to this cluster with a membership degree larger or equal to a given threshold membership degree μ h :…”

“…The identification algorithm is realized through evolving clustering procedure [7] that copes the advantages of the objective function clustering enabling to identify clusters with a generic shape and orientation. It uses Gustafson-Kessel (GK) distance measure [9] to find elliptic clusters with different shape and orientation adapted to cover the individual character of the clustered data.…”

Recursive clustering based on a Gustafson–Kessel algorithm

“…Recently an extension of the Gustafson-Kessel clustering algorithm for evolving data stream clustering was published as a chapter in a book (Filev and Georgieva 2010). The algorithm is also based on a GK off-line clustering algorithm as ours, but the adaptation of centers and the calculation of the fuzzy covariance matrix are different.…”