Marjory C. C. Abreu

Investigating the influence of the choice of the ensemble members in accuracy and diversity of selection-based and fusion-based methods for ensembles

One of the most important steps in the design of a multi-classifier system (MCS), also known as ensemble, is the choice of the components (classifiers). This step is very important to the overall performance of a MCS since the combination of a set of identical classifiers will not outperform the individual members. The ideal situation would be a set of classifiers with uncorrelated errors - they would be combined in such a way as to minimize the effect of these failures. This paper presents an extensive evaluation of how the choice of the components (classifiers) can affect the performance of several combination methods (selection-based and fusion-based methods). An analysis of the diversity of the MCSs when varying their components is also performed. As a result of this analysis, it is aimed to help designers in the choice of the individual classifiers and combination methods of an ensemble.

Investigating the Use of an Agent-Based Multi-classifier System for Classification Tasks

This paper proposes NeurAge, an agent-based multi-classifier system for classification tasks. This system is composed of several neural classifiers (called neural agents) and its main aim is to overcome some drawbacks of multi-classifier systems and, as a consequence, to improve performance of such systems.

Performance and diversity evaluation in hybrid and non-hybrid structures of ensembles

This paper presents a wide evaluation of performance and diversity in hybrid and non-hybrid structures of ensembles. In applying some diversity measures at the chosen ensemble, it is intended to analyse the effect of varying diversity in ensembles and how the variation of diversity can affect the performance of several combination methods (selection-based and combination-based methods). Finally, it is also intended to understand the reasons that some combination methods are more affected by variation in diversity.

An analysis of data distribution in the ClassAge system: An agent-based system for classification tasks

The ClassAge (classifier agents) system has been proposed as an alternative to transform the centralized decision-making process of a multi-classifier system into a distributed, flexible and incremental one. This system has presented good results in some conventional (centralized) classification tasks. Nevertheless, in some classification tasks, relevant features might be distributed over a set of agents. These applications can be classified as distributed classification tasks and a method for distributing data (features or attributes) among the agents is needed. In this paper, an investigation of the impact of using data distribution among the agents in the performance of ClassAge will be performed. In this investigation, the performance of the ClassAge system will be compared with some existing multi-classifier systems. In all combination systems, a feature distribution method based on the Pearson correlation will be used.

Using weighted dynamic classifier selection methods in ensembles with different levels of diversity

There are two main approaches to combine the output of classifiers within a multi-classifier system (MCS), which are: combination-based and selection-based methods. In selection-based methods, only one classifier is needed to correctly classify the input pattern. The choice of a classifier is typically based on the certainty of the current decision. The use of weights can be very useful for the final decision of a selection-based MCS since it can provide a confidence degree for each classifier. This paper presents the use of two confidence measures applied in three selection-based methods. The main aim of this paper is to analyze the benefits of using weights in the main selection-based methods and which confidence measure is more suitable to be used.

Investigating the use of a multi-agent system for knowledge discovery in databases

This paper presents an agent-based system which is capable of performing knowledge discovery from database, including image databases. The proposed system is composed of several knowledge discovery (KD) agents which are responsible for data preparation and data mining. In the system, each agent has its own data mining method and they can have different answers to the same input pattern. In this sense, a negotiation method was proposed. As a result of this method, a commonly agreed output for an input pattern is provided by the system.

Using fuzzy, neural and fuzzy-neural combination methods in ensembles with different levels of diversity

Classifier Combination has been investigated as an alternative to obtain improvements in design and/or accuracy for difficult pattern recognition problems. In the literature, many combination methods and algorithms have been developed, including methods based on computational Intelligence, such as: fuzzy sets, neural networks and fuzzy neural networks. This paper presents an evaluation of how different levels of diversity reached by the choice of the components can affect the accuracy of some combination methods. The aim of this analysis is to investigate whether or not fuzzy, neural and fuzzy-neural combination methods are affected by the choice of the ensemble members.

A Neuro-Fuzzy-Based Agent System with Data Distribution among the Agents for Classification Tasks

The use of intelligent agents in the structure of multi-classifier systems has been investigated in order to overcome some drawbacks of these systems. As a result of this, the NeurAge system was proposed, which is a Neuro-based multi-agent system. This system has presented good results in some conventional (centralized) and distributed classification tasks. In this paper, instead of neural networks, the NeurAge agents will be composed of neuro-fuzzy networks. The main aim of this paper is to investigate the benefits of using fuzzy concepts in a neural-based multi-agent system in which data are distributed among its agents.

Aplicação de Multiclassificadores Heterogêneos no Reconhecimento de Classes Estruturais de Proteínas

O reconhecimento de dobras de proteina e um dos principais problemas em aberto da biologia molecular e uma importante abordagem para a descoberta de estruturas de proteinas desconsiderando a similaridade de suas sequencias. Neste contexto, as ferramentas computacionais, principalmente as tecnicas da Aprendizagem de Maquina (AM), tornaram-se alternativas essenciais para tratar esse problema, considerando o grande volume de dados empregado. Este trabalho apresenta os resultados obtidos com a aplicacao de diferentes sistemas multiclassificadores heterogeneos (Stacking, StackingC e Vote), empregando tipos distintos de classificadores base (Arvores de Decisao, K-Vizinhos Mais proximos, Naive Bayes, Maquinas de Vetores Suporte e Redes Neurais), a tarefa de predicao de classes estruturais de proteina.

Uma análise comparativa de sistemas de classificação usando uma base de estrutura de proteínas

Este artigo apresenta uma análise comparativa de alguns sistemas de classificação aplicadas ao reconhecimento de estruturas de proteínas. Entre os sistemas analisados, serão investigados alguns métodos de negociação aplicados ao sistema NeurAge, que é um sistema multi-agentes composto de vários agentes classificadores. O principal objetivo deste sistema é superar alguns problemas encontrados nos sistemas multiclassifciadores e, como conseqüência, melhorar o desempenho de tais sistemas.