Empirical evidence shows that ensembles with adequate levels of pairwise diversity among a set of accurate member algorithms significantly outperform any of the individual algorithms. As a result, several diversity measures have been developed for use in optimizing ensembles. We show that diversity measures that properly combine the diversity space in an additive and multiplicative manner, not only result in ensembles whose accuracy is comparable to the naive ensemble of choosing the most accurate learners, but also results in ensembles that are significantly more efficient than such naive ensembles. In addition to diversity measures found in the literature, we submit two measures of diversity that span the diversity space in unique ways. Each of these measures considers not only the diversity of ratings between a pair of algorithms, but how this diversity relates to the target values.

Orientador: Hildete Prisco Pinheiro / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matematica, Estatistica e Computação Cientifica / Made available in DSpace on 2018-08-10T18:03:10Z (GMT). No. of bitstreams: 1 Maia_RafaelPimentel_M.pdf: 1076568 bytes, checksum: b09cf45158c8c66c120c29b2addafe8c (MD5) Previous issue date: 2008 / Resumo: O objetivo deste trabalho é propor novas metodologias para avaliar o desempenho dos alunos da UNICAMP, do ingresso à conclusão do curso. O conjunto de dados disponível foi obtido a partir dos questionários Sócio-Culturais aplicados pela Comissão Permanente de Vestibulares (COMVEST) na inscrição do vestibular e informações acadêmicas fornecidas pela Diretoria Acadêmica (DAC) da UNICAMP. Estes se referem às informações de todos os alunos ingressantes nos anos de 1997 a 2000. São propostas duas metodologias, uma com base na variável denominada "ganho relativo" sugerido por Dachs e Maia (2006) e a segunda utilizando as notas de todas as disciplinas cursadas pelos alunos durante a graduação. Essas novas metodologias baseiam-se em medidas de diversidades propostas por Rao (1982) e na utilização de U-Estatísticas. São propostos testes de homogeneidade para avaliar se existe diferença no desempenho entre alunos de grupos distintos (alunos oriundos de escola pública ou privada, por exemplo). Aspectos teóricos de U-Estatística e medidas de diversidade também são apresentados. Para a primeira metodologia foram feitas duas abordagens: paramétrica e não paramétrica, enquanto que para a segunda, apenas a abordagem não paramétrica foi explorada. Na abordagem paramétrica as estimativas são feitas por máxima verossimilhança e na não paramétrica foi utilizado o método de re-amostragem por jackknafe para se obter as estimativas das variâncias. Todas as aplicações utilizaram os dados dos alunos ingressantes / Abstract: The main interest of this work is to propose new methods to evaluate the performances of the students at UNICAMP from admission to graduation. The data was obtained from questionnaires applied by the University Commission of admission's exam (COMVEST) during registration of the exam and academic informations provided by the Directory of Academic Studies (DAC). The data refer to information with respect to all the students enrolled in the University from 1997 to 2000. We propose two methods: one based on the variable "relative gain"(Dachs and Maia, 2006) and the other method uses information about the grades of all courses attended by the students during their undergraduate studies. These new methods are based on diversity measures proposed by Rao (1982) and the use of U-Statistics. Homogeneity tests are proposed to evaluate differences in the performance of the students according to different socio-economic groups. For the first method, we have two approaches: a parametric and a nonparametric analysis. For the second method, only a nonparametric analysis was done. In the parametric analysis, a Maximum Likelihood Estimation procedure is used and in the nonparametric analysis, resampling methods such as jackknife was used to obtain the estimates of the variances and confidence intervals. All the applications use the data of the enrolled students / Mestrado / Probabilidade e Estatistica Aplicada / Mestre em Estatística

The global loss of biodiversity calls for robust large-scale diversity assessment. Biological diversity is a multi-faceted concept; defined as the “variety of life”, answering questions such as “How much is there?” or more precisely “Have we succeeded in reducing the rate of its decline?” is not straightforward. While various aspects of biodiversity give rise to numerous ways of quantification, we focus on temporal (and spatial) trends and their changes in species diversity. Traditional diversity indices summarise information contained in the species abundance distribution, i.e. each species' proportional contribution to total abundance. Estimated from data, these indices can be biased if variation in detection probability is ignored. We discuss differences between diversity indices and demonstrate possible adjustments for detectability. Additionally, most indices focus on the most abundant species in ecological communities. We introduce a new set of diversity measures, based on a family of goodness-of-fit statistics. A function of a free parameter, this family allows us to vary the sensitivity of these measures to dominance and rarity of species. Their performance is studied by assessing temporal trends in diversity for five communities of British breeding birds based on 14 years of survey data, where they are applied alongside the current headline index, a geometric mean of relative abundances. Revealing the contributions of both rare and common species to biodiversity trends, these "goodness-of-fit" measures provide novel insights into how ecological communities change over time. Biodiversity is not only subject to temporal changes, but it also varies across space. We take first steps towards estimating spatial diversity trends. Finally, processes maintaining biodiversity act locally, at specific spatial scales. Contrary to abundance-based summary statistics, spatial characteristics of ecological communities may distinguish these processes. We suggest a generalisation to a spatial summary, the cross-pair overlap distribution, to render it more flexible to spatial scale.

No description available.