Authors: Leila Maria Vriesmann; Alceu De Souza Britto
Addresses: State University of Ponta Grossa, Ponta Grossa, PR, Brazil Pontifical Catholic University of Parana, Curitiba, PR, Brazil ' State University of Ponta Grossa (UEPG), Av. General Carlos Cavalcanti, 4748, Ponta Grossa (PR), 84030-900, Brazil; Pontifical Catholic University of Parana (PUCPR), R. Imaculada Conceição, 1155, Curitiba (PR), 80215-901, Brazil ' Federal University of Parana (UFPR), Av. Cel. Francisco H. dos Santos, s/n, Curitiba (PR), 81530-900, Brazil. ' École de Technologie Supérieure (ÉTS) – University of Quebec, 1100, Notre-Dame West, Montréal (QC), H3C1K3, Canada. ' École de Technologie Supérieure (ÉTS) – University of Quebec, 1100, Notre-Dame West, Montréal (QC), H3C1K3, Canada
Abstract: This work evaluates some strategies to approximate the performance of a dynamic ensemble selection method to the oracle performance of its pool of weak classifiers. For this purpose, we evaluated different distance metrics in the K-nearest-oracles (KNORA) method, the use of statistics related to the class accuracy of each classifier in the pool and some additional information calculated by using a clustering process in the validation dataset. Moreover, different strategies are also evaluated to combine the results of the KNORA dynamic ensemble selection method with the results of its built-in K-nearest neighbour (KNN) used to define the neighbourhood of a test pattern during the ensemble creation. A strong experimental protocol based on more than 60,000 samples of handwriting digits extracted from NIST-SD19 was used to evaluate each strategy. The experiments have shown that the fusion of the KNORA results with the results of its built-in KNN is a very promising strategy.
Keywords: dynamic ensemble selection; oracle information; K-nearest-oracles; KNORA; distance metrics; handwriting samples.
International Journal of Innovative Computing and Applications, 2012 Vol.4 No.3/4, pp.184 - 200
Available online: 26 Oct 2012 *Full-text access for editors Access for subscribers Purchase this article Comment on this article