English

CIn - Centro de Informática UFPE




Eventos Relacionados

Defesa de Dissertação de Mestrado Nº 1.251: "An Authomatic Method for Construction of Multi-classifier Systems Based on the Combination of Selection and Fusion"

O aluno Tiago Pessoa Ferreira de Lima irá defender seu trabalho dia 26 de fevereiro, às 10h, na Sala D224 Início: 26/02/2013 às 10:00 Término: 26/02/2013 às 12:00 Local: Sala D224

Pós-Graduação em Ciência da Computação – UFPE
Defesa de Dissertação de Mestrado Nº 1.251
 
Aluno: Tiago Pessoa Ferreira de Lima
Orientadora: Profa. Teresa Bernarda Ludermir
Título: AN AUTOMATIC METHOD FOR CONSTRUCTION OF MULTI-CLASSIFIER SYSTEMS BASED ON THE COMBINATION OF SELECTION AND FUSION
Data: 26/02/2013
Hora/Local: 10:00h – Sala  D224
Banca Examinadora:
Prof. Leandro Maciel Almeida   (UFPE / CIn)
Profa. Clarissa Daisy da Costa Albuquerque  (UNICAP / Dep. de Estatística e Informática)
Profa. Teresa Bernarda Ludermir  (UFPE / CIn)
 
RESUMO:
 
In this master thesis, we present a methodology that aims to automatic construction of multi-classifiers systems based on the combination of selection and fusion. The proposed method initially finds the optimum number of clusters for training data set and subsequently determines an ensemble for each cluster found. For model evaluation, the testing data set are submitted to clustering technique and the nearest cluster to data input will emit a supervised response through its associated ensemble. In current work, k-means and self-organizing maps were used in clustering phase, and multilayer Perceptron was used in the classification phase. Adaptive Differential Evolution has been used in this work in order to optimize the parameters and performance of the different techniques used in classification and clustering phases. The proposed method, called SFJADE, has been tested on 7 benchmark problems in machine learning and neural networks, including cancer, card, diabetes, glass, heart, heartc and horse. The experimental results have show that the SFJADE method has better performance than some literature methods, besides significantly outperforms the most of the methods commonly used to construct Multi-Classifier Systems.
 
Palavras-chave: Multi-classifier systems, ensembles, selection and fusion, k-means, self-organizing maps, multilayer perceptron, adaptive differential evolution. 
  • © Centro de Informática UFPE - Todos os direitos reservados
    Tel +55 81 2126.8430 - Cidade Universitária - 50740-560 - Recife/PE
Plano4 Consultoria Web