Abstract:Multi-classifier systems (MCSs) are some kind of predictive models that classify instances by combining the output of an ensemble of classifiers given in a pool. With the aim of enhancing the performance of MCSs, dynamic selection (DS) techniques have been introduced and applied to MCSs. Dealing with each test sample classification, DS methods seek to perform the task of classifier selection so that only the most competent classifiers are selected. The principal subject regarding DS techniques is how the competence of classifiers corresponding to every new test sample classification task can be estimated. In traditional dynamic selection methods, for classifying an unknown test sample x, first, a local region of data that is similar to x is detected. Then, those classifiers that efficiently classify the data in the local region are also selected so as to perform the classification task for x. Therefore, the main effort of these methods is focused on one of the two following tasks: (i) to provide a measure for identifying a local region, or (ii) to provide a criterion for measuring the efficiency of classifiers in the local region (competence measure). This paper proposes a new version of dynamic selection techniques that does not follow the aforementioned approach. Our proposed method uses a multi-label classifier in the training phase to determine the appropriate set of classifiers directly (without applying any criterion such as a competence measure). In the generalization phase, the suggested method is employed efficiently so as to predict the appropriate set of classifiers for classifying the test sample x. It is remarkable that the suggested multi-label-based framework is the first method that uses multi-label classification concepts for dynamic classifier selection. Unlike the existing meta-learning methods for dynamic ensemble selection in the literature, our proposed method is very simple to implement and does not need meta-features. As the experimental results indicate, the suggested technique produces a good performance in terms of both classification accuracy and simplicity which is fairly comparable with that of the benchmark DS techniques. The results of conducting the Quade non-parametric statistical test corroborate the clear dominance of the proposed method over the other benchmark methods.

Design of Dynamic Multiple Classifier Systems Based on Belief Functions

A Multi-Class Large Margin Classifier

Combination of Heterogeneous Multiple Classifiers Based on Evidence Theory

Dynamic Base Classifier Pool for Classifier Selection in Multiple Classifier Systems.

A survey of multiple classifier systems as hybrid systems

Multiple Classifier Fusion Based on Testing Sample Pairs

Dynamic fusion method using Localized Generalization Error Model

Design Of Multiple Classifier Systems Based On Evidential Neural Network

Multiple K-Nn Classifiers Fusion Based on Evidence Theory

Experimental Comparison Between Implicit and Explicit MCSs Construction Methods

A Novel Diversity Measure of Multiple Classifier Systems Based on Distance of Evidence

Knowledge discovery-based multiple classifier fusion: a generalized rough set method

Multi-Class Svm Classifiers Fusion Based On Evidence Combination

Multiple Classifiers Fusion Based On Weighted Evidence Combination

Design of a multiple classifier system

A New Diversity Measure for Classifier Fusion

A novel framework based on the multi-label classification for dynamic selection of classifiers

Design Of Multiple Classifier Systems Based On Sfls

A novel dynamic fusion method using localized generalization error model

Dynamic Classifier Ensemble Using Classification Confidence

MFA-DMFS:new Multi-Classifier Fusion Method and Its Application