Distance-based online classifiers. (30th October 2016)

• Record Type: Journal Article
• Authors: Jędrzejowicz, Joanna; Jędrzejowicz, Piotr
• Is Part Of: Expert systems with applications. Volume 60(2016)
• Page Start: 249
• Page End: 257
• Publication Date: 2016-10-30
• Subjects: Online learning -- Fuzzy C-means clustering -- Kernelized clustering -- Rotation forest; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2016.05.015
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: Proposing the online distance-based classifier with fuzzy C-means clustering. Evaluating different strategies for online updating training dataset. Adopting Rotation Forest method to extend the online distance-based classifier. Analysis of the computational complexity of the proposed algorithms. Extensive computational experiment validating the proposed algorithms. Abstract: Main impact of the paper is proposing a family of algorithms for the online learning and classification. These algorithms work in rounds, where at each round a new instance is given and the algorithm makes a prediction. After the true class of the instance is revealed, the learning algorithm updates its internal hypothesis. The proposed algorithms are based on fuzzy C-means clustering and kernel-based fuzzy C-means clustering, followed by a calculation of distances between cluster centroids and the incoming instance for which the class label is to be predicted. In one of the proposed variants, simple distance-based classifiers thus obtained serve as basic classifiers for the implemented Rotation Forest ensemble classifier, which increases the accuracy of classification. In the paper we also propose using kernelized fuzzy C-means clustering method as an alternative approach to constructing distance based online classifiers. The approach allows to construct online classifiers of the polynomial computational complexity which is a significant feature considering potential application to the big data analysis. Using the kernelized clustering is advantageous since it allows for automatic estimation of the number of clusters maintaining the number of the user-defined parameters. The proposed classification algorithms are validated experimentally. Experiment results show that the approach assures good quality of classification, extending the range of the available online approaches.