Uncertain fuzzy self-organization based clustering: interval type-2 fuzzy approach to adaptive resonance theory
Introduction
In most unsupervised clustering algorithms, the desired number of clusters may not be estimated efficiently by the algorithm itself, since this value is required as an input parameter, such as in K-means clustering, or computation of complex statistics must be performed for this decision, such as the CH-index [2] and gap statistic [45] for hierarchical clustering. For this reason, various methods have been proposed to achieve self-organizing clustering, such as the adaptive resonance theory (ART) [4].
ART networks have emerged as intelligent and autonomous learning algorithms due to the following properties: self-organization, self-stabilization, plasticity, and real-time learning, where numerous network variants have been proposed, such as fuzzy ART [6], ARTMAP [5], ARTMAP-pi, dART, dARTMAP [3], and falcon ART [25], to name a few.
Although ART-based methods may provide a satisfactory baseline solution to the problem of assigning dynamic number and size of clusters, their performance is still restricted due to the requirement of selecting an empirical vigilance parameter value. For this reason, integrating fuzzy sets with ART methods, as in fuzzy ART [6], have proved to be successful in imitating the ambiguity in the number and size of clusters. They have since been utilized in a number of applications in diverse fields such as gene expression analysis [46] and image processing [8].
In order to improve the clustering performance of conventional ART, we propose a method to define the vigilance parameter as a function of the relative distance between the current centers of the candidate cluster and the input vector. Such a definition is based on the radial nature of the clusters present in most real datasets. The proposed vigilance function can then be viewed as a type-1 (T1) fuzzy membership function, where for each incoming pattern a suitable vigilance value may be obtained.
For any pattern recognition application, it may not always be possible to extract perfect knowledge from a pattern set. This ambiguity may lead to an uncertain choice of parameters values to represent data that may result in the formation of inappropriate prototypes. To mitigate this problem, many researchers have applied type-2 (T2) fuzzy methods to several applications, for example, medical diagnosis [18], community transport scheduling [19], reactive robot navigation [14], [26], chemical analysis [50], and survey processing [20], to name a few. Accordingly, the uncertainty associated with the vigilance parameter may be managed by extending it to a T2 fuzzy set (FS). In general, T2 FSs are capable of modelling various uncertainties that cannot be properly managed by T1 FSs, at the cost of added computational complexity.
To reduce this complexity, interval type-2 (IT2) FSs were proposed, since all secondary grades in these FSs are uniformly weighted (i.e., all equal to one). These MFs are also convenient for defuzzification, since the T1 fuzzy MF obtained after type reduction is an interval, which may be characterized by two fixed values. In light of this, we propose an IT2 fuzzy ART approach to further improve the performance of the proposed T1 fuzzy ART using the adaptive vigilance method, by incorporating the associated uncertainty of the data using upper and lower vigilance membership functions.
Our key contributions in this paper may be summarized as follows.
- •
We propose novel methods (T1 fuzzy ART with adaptive vigilance and IT2 fuzzy ART approach) for dynamic clustering that also takes into account the uncertainty associated with clusters in most real world applications.
- •
We model the vigilance parameter in ART clustering using T1 fuzzy and IT2 fuzzy MFs, and exploit the KM algorithm for defuzzification. Our method differs from conventional fuzzy ART in that the MF may be selected heuristically depending upon an estimate of the number and size of the clusters.
- •
We argue that the IT2 fuzzy approach may be suitable for extending any clustering algorithm that requires dynamic selection of clusters, due to its inherent ability to model uncertainties.
The remainder of this paper is organized as follows. We provide a brief review of the related research in Section 2. In Section 3, we discuss the conventional fuzzy ART algorithm and the effect of the vigilance parameter on the network dynamics. We also summarize the concepts of IT2 fuzzy sets, type reduction, and defuzzification. In Section 4, we describe our proposed methods, namely T1 fuzzy ART and IT2 fuzzy ART. Experimental results showing the effectiveness and improved performance of the algorithms are presented in Section 5. We conclude with a discussion of future work in Section 6.
Section snippets
Related research
Previously, complex fuzzy methods for adapting the vigilance parameter for ART-II network have been proposed for telecommunication signals [24]. However, the converging time for this method can be slow and requires specification of an empirical value of the initial vigilance that may affect the performance accordingly. In a previous work, evolutionary computation techniques were used to automate the selection of parameters for fuzzy ART [23]. However, no criterion has been specified to stop the
Fuzzy ART algorithm
The fundamental feature of all ART systems is a matching process that results in either of two states: a resonant state that focuses attention and triggers stable prototype learning, or a self-regulating parallel memory search [6]. The architecture of the fuzzy ART network, shown in Fig. 1, consists of two subsystems: an attentional subsystem and an orienting subsystem.
- 1.
The attentional subsystem consists of an input layer F0, which receives an M-dimension input vector, a membership computing
Proposed methods
In this section, we describe our proposed methods, namely T1 fuzzy ART with adaptive vigilance and IT2 fuzzy ART clustering.
Experimental results
In this section, we illustrate various experimental results to compare the clustering performance of the conventional fuzzy ART algorithm, the proposed T1 fuzzy ART with adaptive vigilance, and the IT2 fuzzy ART algorithm. We begin by demonstrating the clustering procedure for simple synthetic data and then proceed to higher dimensional data sets. We also provide comparisons of the clustering methods for image segmentation problems.
For the conventional fuzzy ART algorithm, experimental results
Conclusion
ART-based clustering methods offer self-organization and simplicity. However, the fixed value of the vigilance parameter ρ in the algorithm limits the flexibility of the cluster generation, which inhibits the performance of the algorithm when the input pattern space consists of clusters of different volumes. In this paper, we proposed methods for modifying the conventional fuzzy ART algorithm by dynamically adapting and selecting the value of the vigilance parameter ρ for each pattern, as a
Acknowledgment
This work was supported by the Technology Innovation Program of the Korea Institute for Advancement of Technology (KTA) granted financial resource from the Ministry of Trade, Industry & Energy, Republic of Korea (No. 2015-122).
References (50)
- et al.
An efficient approach for electric load forecasting using distributed ART (adaptive resonance theory) & HS-ARTMAP (hyper-spherical ARTMAP network) neural network
Energy
(2011) Distributed learning, recognition, and prediction by ART and ARTMAP neural networks
Neural Netw.
(1997)- et al.
A massively parallel architecture for a self-organizing neural pattern recognition machine
Comput. Vis. Graph. Image Process.
(1987) - et al.
ARTMAP: supervised real-time learning and classification of nonstationary data by a self-organizing neural network
Neural Netw.
(1991) - et al.
Fuzzy ART: fast stable learning and categorization of analog patterns by an adaptive resonance system
Neural Netw.
(1991) - et al.
Interval type-2 fuzzy membership function generation methods for pattern recognition
Inf. Sci.
(2009) - et al.
Centroid of a type-2 fuzzy set
Inf. Sci.
(2001) - et al.
A passport recognition and face verification using enhanced fuzzy ART based RBF network and PCA algorithm
Neurocomputing
(2008) - et al.
A review on type-2 fuzzy logic applications in clustering, classification and pattern recognition
Appl. Soft. Comput.
(2014) - et al.
A comparative experimental study of type-1/type-2 fuzzy cascade controller based on genetic algorithms and particle swarm optimization
Expert Syst. Appl.
(2011)
Interval-valued fuzzy set approach to fuzzy co-clustering for data classification
KnowBased Syst. 107
Maximum certainty data partitioning
Pattern Recognit.
Fuzzy kohonen clustering networks
Pattern Recognit.
An efficient k-means clustering algorithm
Pattern Recognit. Lett.
A dendrite method for cluster analysis
Commun. Stat. Theor. Methods
Application of interval type-2 fuzzy neural networks in non-linear identification and time series prediction
Soft. comput.
Fuzzy-ART based adaptive digital watermarking scheme
IEEE Trans. Circuits Syst. Video Technol.
Fuzzy clustering and aggregation of relational data with instance-level constraints
IEEE Trans. Fuzzy Syst.
Mercer kernel-based clustering in feature space
IEEE Trans. Neural Netw.
A hierarchical type-2 fuzzy logic control architecture for autonomous mobile robots
IEEE Trans. Fuzzy Syst.
An improved learning rule for fuzzy ART
J. Inf. Sci. Eng.
Uncertain fuzzy clustering: interval type-2 fuzzy approach to c-means
IEEE Trans. Fuzzy Syst.
Cited by (18)
Robust interval type-2 kernel-based possibilistic fuzzy clustering algorithm incorporating local and non-local information
2023, Advances in Engineering SoftwareInterval type-2 fuzzy logic system based similarity evaluation for image steganography
2020, HeliyonCitation Excerpt :Hence, a significant improvement has been made in shifting from a type-1 fuzzy logic system (T1 FLS) to interval type-2 fuzzy logic system (IT2 FLS) by the researchers in recent years. IT2 FLSs have been effectively applied in the various applications of image processing systems such as classification (Majeed et al., 2018; Rubio et al., 2017), filtering (Singh et al., 2018), segmentation (Dhar and Kundu, 2019; Zhao et al., 2019), and edge detection (Castillo et al., 2017; Gonzalez and Melin, 2017; Gonzalez et al., 2016; Martínez et al., 2019; Melin et al., 2014). In this paper, we propose a novel steganographic procedure using an IT2 FLS based similarity measure to measure the similarity between the pixels in a digital image.
A survey of adaptive resonance theory neural network models for engineering applications
2019, Neural NetworksCitation Excerpt :However, it is often set empirically in an ad hoc manner. In the unsupervised learning mode, vigilance adaptation has been addressed in fuzzy ART through the usage of game theory (Fudenberg & Tirole, 1991) in Fung and Liu (1999); the activation maximization, confliction minimization and hybrid integration rules in Meng, Tan, and Wunsch II (2013, 2019) and Meng et al. (2016); the combination with particle swarm optimization (Kennedy & Eberhart, 1995) and cluster validity indices (Xu & Wunsch II, 2009) in Smith and Wunsch II (2015); defining the vigilance as a function of the category size in Isawa et al. (2008b, 2009); or modeling it as a fuzzy membership function in Majeed et al. (2018). Despite these contributions, setting the vigilance parameter still remains a challenging task worthy of further exploration, particularly in the online learning mode.
Towards interval-valued fuzzy set-based collaborative fuzzy clustering algorithms
2018, Pattern RecognitionCitation Excerpt :Besides, several type-2 fuzzy set-based methods have been studied for data classification [44,45] in which an interval type-2 fuzzy logic system was used for gene selection and cancer classification; or to posibilistic type-2 fuzzy clustering [38] and to design the type-2 fuzzy logic systems using type-2 fuzzy clustering [40]. Recently, Majeed et al. [50]. proposed fuzzy adaptive resonance theory (ART) method that adaptively selects the vigilance parameter value and then be extended to an interval type-2 fuzzy ART method to handle the uncertainty associated with memberships.
Radar signal recognition based on deep convolutional neural network in complex electromagnetic environment
2022, 3rd China International SAR Symposium, CISS 2022Typical Characteristic-Based Type-2 Fuzzy C-Means Algorithm
2021, IEEE Transactions on Fuzzy Systems