Subtractive Clustering and Particle Swarm Optimization Based Fuzzy Classifier

Subtractive Clustering and Particle Swarm Optimization Based Fuzzy Classifier

Halima Salah, Mohamed Nemissi, Hamid Seridi, Herman Akdag
Copyright: © 2019 |Pages: 15
DOI: 10.4018/IJFSA.2019070105
(Individual Articles)
No Current Special Offers


Setting a compact and accurate rule base constitutes the principal objective in designing fuzzy rule-based classifiers. In this regard, the authors propose a designing scheme based on the combination of the subtractive clustering (SC) and the particle swarm optimization (PSO). The main idea relies on the application of the SC on each class separately and with a different radius in order to generate regions that are more accurate, and to represent each region by a fuzzy rule. However, the number of rules is then affected by the radiuses, which are the main preset parameters of the SC. The PSO is therefore used to define the optimal radiuses. To get good compromise accuracy-compactness, the authors propose using a multi-objective function for the PSO. The performances of the proposed method are tested on well-known data sets and compared with several state-of-the-art methods.
Article Preview


Fuzzy approaches have become one of the well-known solutions for the classification problems due to their powerful capabilities to handle both imprecision and uncertainty concepts. The generation of membership function and the generation of the associate fuzzy rules constitute the key tasks in the design of fuzzy systems. For this purpose, a large variety of methods have been introduced. This includes: heuristic methods (Abe & Lan, 1995; Nozaki et al., 1997; Viattchenin et al., 2013), Genetic Algorithms (GA) (Antonelli et al., 2014;Cordon et al., 1998; Mansoori et al., 2008; Yuan & Zhuang, 1996), Artificial Immune Systems (AIS) (Alves et al., 2004; Lei & Ren-hou, 2008), Ant Colony Optimization (ACO) (Ganji & Abadeh, 2011; Saniee et al., 2008), Particle Swarm Optimization (PSO) (Chen, 2006; Ganeshkumar et al., 2013; Kashanipour et al., 2008; Permana & Hashim, 2010), Neural Networks (Chen et al., 1995; Nemissi et al., 2014; Pranevicius et al., 2014), and clustering methods (Chen et al., 2011; Jamsandekar & Mudholkar, 2013; Jamsandekar & Mudholkar, 2014; Kahali et al., 2017; Mahata et al., 2018).

The advantage of clustering based-approaches, in which every rule corresponds to one cluster, lies in the fact that they permit generating efficient classifiers with few rules. Indeed, in grid-partitions methods, the feature space is divided into all possible regions and each of them is represented by one fuzzy rule. This process generates a large number of rules, especially in high-dimensional problems where the number of rules becomes exponentially huge (curse of dimensionality). These rules have to be then selected and tuned. Furthermore, in clustering based-approaches the parameters of the membership functions, i.e. centers and width, can be automatically set by projecting the obtained clusters.

Complete Article List

Search this Journal:
Volume 13: 1 Issue (2024)
Volume 12: 1 Issue (2023)
Volume 11: 4 Issues (2022)
Volume 10: 4 Issues (2021)
Volume 9: 4 Issues (2020)
Volume 8: 4 Issues (2019)
Volume 7: 4 Issues (2018)
Volume 6: 4 Issues (2017)
Volume 5: 4 Issues (2016)
Volume 4: 4 Issues (2015)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing