Generalized Net Model Simulation of Cluster Analysis Using CLIQUE: Clustering in Quest

摘要

Cluster analysis searches for similarities between data objects according to their characteristics and groups the similar objects into clusters. One of the techniques which combines subspace grid-based clustering and density-based cluster analysis, namely Clustering In Quest (CLIQUE), is studied in the present research. The main steps performed in the process of detecting groups of objects with similar behaviour are: dividing the data space into a finite number of cells, forming a grid-based structure, detecting groups of similar objects and defining the clusters. Generalized Nets (GNs) have been introduced by Atanassov as an extension of the ordinary Petry nets and other their extensions and modifications. They are a powerful tool for modelling real processes. A GN-model of the CLIQUE real-time data clustering process is constructed here and a simulation of the model is performed using a platform independent software, called GN Integrated Development Environment (GN IDE). An open-source version of the RapidMiner software is used for performing the cluster analysis on real datasets.