Processamento eficiente de padrões de composição geométrica em grandes volumes de dados

• Autor(a) principal: Khatibi, Amir Hassan Moghadam
• Data de Publicação: 2016
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações do LNCC
• Texto Completo: https://tede.lncc.br/handle/tede/284
• topic: Big data; Indexação espacial; Amostragem; Processamento eficiente; Processamento paralelo; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::METODOLOGIA E TECNICAS DA COMPUTACAO::BANCO DE DADOS

Big data processing is expected to empower decision-making as more infor- mation becomes accessible to analytical tools. In this work, we argue that the data deluge produced by the Big Data phenomenon blurs, amongst billions of dataset elements, high-level objects that can only be perceived once adequate composition models are in place. We argue that identifying such objects is relevant for various disciplines and we focus on Geometric Composition model using an example in astronomy. This work formulates the problem of Efficiently Computing Geometric Composition Patterns in Big Data (GCP). We present a novel technique with some pruning strategies to find these objects while we show the robustness and efficiency of our technique using a SPARK implementation over Hadoop parallel architecture.