Browsing by Author "Almeida, Humberto Mossri de"

Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Modelo de gestão de demandas de manutenção de software : a experiência da PUC Minas.
    (2009) Malta, Marcelo Nassau; Almeida, Humberto Mossri de; Valente, Marco Túlio de Oliveira; Pietrobon, Carlos Alberto Marques; Marques Neto, Humberto Torres
    Em um contexto no qual a demanda é maior do que a capacidade de atendimento gerenciar serviços e recursos é essencial. Neste trabalho, apresenta-se um modelo de gestão de solicitações de manutenção de software para planejar e organizar a demanda dos usuários, bem como incentivar o atendimento por projetos e reduzir o volume de atendimentos em caráter emergencial. Para avaliar a eficácia do modelo proposto, foram comparados períodos similares antes e depois da implantação no ambiente da Pontifícia Universidade Católica de Minas Gerais (PUC Minas).
  • No Thumbnail Available
    Item
    WCL2R : a benchmark collection for Learning to rank research with clickthrough data.
    (2010) Alcântara, Otávio D. A.; Pereira Junior, Álvaro Rodrigues; Almeida, Humberto Mossri de; Gonçalves, Marcos André; Middleton, Christian; Yates, Ricardo Baeza
    WCL2R: A benchmark collection for Learning to rank research with clickthrough data In this paper we present WCL2R, a benchmark collection for supporting research in learning to rank (L2R) algorithms which exploit clickthrough features. Differently from other L2R benchmark collections, such as LETOR and the recently released Yahoo!’s collection for a L2R competition, in WCL2R we focus on defining a significant (and new) set of features over clickthrough data extracted from the logs of a real-world search engine. In this paper, we describe the WCL2R collection by providing details about how the corpora, queries and relevance judgments were obtained, how the learning features were constructed and how the process of splitting the collection in folds for representative learning was performed. We also analyze the discriminative power of the WCL2R collection using traditional feature selection algorithms and show that the most discriminative features are, in fact, those based on clickthrough data. We then compare several L2R algorithms on WCL2R, showing that all of them obtain significant gains by exploiting clickthrough information over using traditional ranking approaches.