Dynamic selection of classiersA comprehensive review Alceu S. Britto Jr. a,b,n , Robert Sabourin c , Luiz E.S. Oliveira d a Pontifícia Universidade Católica do Paraná (PUCPR), Curitiba, PR, Brazil b Universidade Estadual de Ponta Grossa (UEPG), Ponta Grossa, PR, Brazil c École de technologie supérieure (ÉTS), Université du Québec, Montreal, QC, Canada d Universidade Federal do Paraná (UFPR), Curitiba, PR, Brazil article info Article history: Received 28 August 2013 Received in revised form 3 May 2014 Accepted 7 May 2014 Keywords: Ensemble of classiers Dynamic selection of classiers Data complexity abstract This work presents a literature review of multiple classier systems based on the dynamic selection of classiers. First, it briey reviews some basic concepts and denitions related to such a classication approach and then it presents the state of the art organized according to a proposed taxonomy. In addition, a two-step analysis is applied to the results of the main methods reported in the literature, considering different classication problems. The rst step is based on statistical analyses of the signicance of these results. The idea is to gure out the problems for which a signicant contribution can be observed in terms of classication performance by using a dynamic selection approach. The second step, based on data complexity measures, is used to investigate whether or not a relation exists between the possible performance contribution and the complexity of the classication problem. From this comprehensive study, we observed that, for some classication problems, the performance contribution of the dynamic selection approach is statistically signicant when compared to that of a single-based classier. In addition, we found evidence of a relation between the observed performance contribution and the complexity of the classication problem. These observations allow us to suggest, from the classication problem complexity, that further work should be done to predict whether or not to use a dynamic selection approach. & 2014 Elsevier Ltd. All rights reserved. 1. Introduction Classication is a fundamental task in Pattern Recognition, which is the main reason why the past few decades have seen a vast number of research projects devoted to classication methods applied to different elds of the human activity. Although the methods available in the literature may differ in many respects, the latest research results lead to a common conclusion; creating a monolithic classier to cover all the variability inherent to most pattern recognition problems is somewhat unfeasible. With this in mind, many researchers have focused on Multiple Classier Systems (MCSs), and consequently, many new solutions have been dedicated to each of the three possible MCS phases: (a) generation, (b) selection, and (c) integration, which are represented in Fig. 1. In the rst phase, a pool of classiers is generated; in the second phase, one or a subset of these classiers is selected, while in the last phase, a nal decision is made based on the prediction(s) of the selected classier(s). It is worth noting that such a representation is not unique, since the selection and integration phases may be facultative. For instance, one may nd MCS where the whole pool of classiers is used without any selection or systems where just one classier is selected from the pool, making the integration phase unnecessary. In a nutshell, recent contributions with respect to the rst phase indicate that the most promising direction is to generate a pool of accurate and diverse classiers. The authors in [1] state that a necessary and sufcient condition for an ensemble of classiers to be more accurate than any of its individual members is for the classiers to be accurate and diverse. Dietterich [2] explains that an accurate classier has an error rate lower than the random guessing on new samples, while two classiers are diverse if they make different errors on new samples. The rationale behind this is that the individual accurate classiers in the pool may compete each other by making different and perhaps complemen- tary errors. As for the selection phase, interesting results have been obtained by selecting specic classiers for each test pattern, which characterizes a dynamic selection of classiers, instead of using the same classier for all of them (static selection). More- over, additional contributions have been observed when ensem- bles are selected instead of just one single classier. In such a case, Contents lists available at ScienceDirect journal homepage: www.elsevier.com/locate/pr Pattern Recognition http://dx.doi.org/10.1016/j.patcog.2014.05.003 0031-3203/& 2014 Elsevier Ltd. All rights reserved. n Corresponding author at: Post-graduate Program in Informatics (PPGIa), Ponti- cal Catholic University of Parana Rua Imaculada Conceição, 1155, Curitiba (PR), 80215-901, Brazil. Tel.: +55 41 3271 1669; fax: +55 41 3271 2121. E-mail addresses: alceu@ppgia.pucpr.br (A.S. Britto Jr.), robert.sabourin@etsmtl.ca (R. Sabourin), lesoliveira@inf.ufpr.br (L.E.S. Oliveira). Please cite this article as: A.S. Britto Jr. et al., Dynamic selection of classiersA comprehensive review, Pattern Recognition (2014), http://dx.doi.org/10.1016/j.patcog.2014.05.003i Pattern Recognition (∎∎∎∎) ∎∎∎∎∎∎