Graded Possibilistic Clustering of Non-stationary Data Streams

Abstract

Multidimensional data streams are a major paradigm in data science. This work focuses on possibilistic clustering algorithms as means to perform clustering of multidimensional streaming data. The proposed approach exploits fuzzy outlier analysis to provide good learning and tracking abilities in both concept shift and concept drift.

Clustering of nonstationary data streams: A survey of fuzzy partitional methods

“…However, this often requires some additional processing. One of the methods described below (Abdullatif, Masulli, Rovetta, & Cabri, ) is based on a variant of possibilistic clustering (Masulli & Rovetta, ) that limits this phenomenon by allowing the user to set the desired balance between competitive behavior (partitional or “probabilistic” clustering) and cooperative behavior (mode seeking or possibilistic clustering). This is beneficial because it avoids the need of an accurate initialization or post‐processing.…”

“…The graded possibilistic c means stream clustering (GPCM stream; Abdullatif, Masulli, Rovetta, & Cabri, ) method uses as its basic clustering model the graded possibilistic c means (GPCM; Masulli & Rovetta, ) adapted to online (or by‐object) operation. It learns from either individual input objects as soon as they arrive, or a sliding window of fixed size that includes the newest object while forgetting the oldest one.…”