Typicality degrees were defined in supervised learning as a tool to build characteristic representatives for data categories. In this paper, an extension of these typicality degrees to unsupervised learning is proposed to perform clustering. The proposed algorithm constitutes a GustafsonKessel variant and makes it possible to identify ellipsoidal clusters with robustness as regards outliers.

in this paper an application of gustafson-kessel clustering algorithm is presented to create a fault detection map (fdm). five post-stack seismic attributes are extracted from a desired seismic time slice related to 3d seismic data of a gas field located in southwest of iran. to find the optimal cluster numbers, two frequently used clustering validity measures, i.e. sc and xb, are used and then...

According to the growth of the Internet technology, there is a need to develop strategies in order to maintain security of system. One of the most effective techniques is Intrusion Detection System (IDS). Clustering which is commonly used to detect possible attacks is one of the branches of unsupervised learning. Fuzzy clustering algorithms play an important role to reduce spurious alarms and I...

This work proposes how to generate a set of fuzzy rules from a data set using a clustering algorithm, the GKPFCM. If we recommend a number of clusters, the GKPFCM identifies the location and the approximate shape of each cluster. These ones describe the relations among the variables of the data set, and they can be expressed as conditional rules such as "If/Then". The GKPFCM provides membership...

Technology has been successfully applied in sports, where biomechanical analysis is one of the most important areas used to raise the performance of athletes. In this context, this paper focuses on swim velocity profile identification using Radial Basis Functions Neural Networks (RBF-NN) trained by the Gustafson-Kessel clustering combined with a novel Dynamic Self-adaptive Multiobjective Harmon...

In this paper, a methodology for design of fuzzy Kalman filter, using interval type-2 models, in discrete time domain, via spectral decomposition experimental data, is proposed. The adopted consists recursive parametric estimation local state space linear submodels filter tracking and forecasting the dynamics inherited to an version Observer/Kalman Filter Identification (OKID) algorithm. partit...