Browsing by Author "Silva, Rodrigo Rocha"

Now showing 1 - 4 of 4
  • No Thumbnail Available
    Item
    Big high-dimension data cube designs for hybrid memory systems.
    (2020) Silva, Rodrigo Rocha; Hirata, Celso Massaki; Lima, Joubert de Castro
    In Big Data cubes with hundreds of dimensions and billions of tuples, the indexing and query operations are a challenge and the reason is the time-space exponential complexity when a full cube is computed. Therefore, solutions based on RAM may not be practical and the solutions based on hybrid memory (RAM and disk) become viable alternatives. In this paper, we propose a hybrid approach, named bCubing, to index and query high-dimension data cubes with high number of tuples in a single machine and using RAM and disk memory systems. We evaluated bCubing in terms of runtime and memory consumption, comparing it with the Frag-Cubing, HIC and H-Frag approaches. bCubing showed to be faster and used less RAM than Frag-Cubing, HIC and H-Frag. bCubing indexed and allowed to query a data cube with 1.2 billion tuples and 60 dimensions, consuming only 84 GB of RAM, which means 35% less memory than HIC. The complex holistic measures mode and median were computed in multidimensional queries, and bCubing was, on average, 50% faster than HIC.
  • No Thumbnail Available
    Item
    Dynamic topic herarchies and segmented rankings in textual OLAP technology.
    (2017) Souza, Adriano Neves de Paula e; Lima, Joubert de Castro; Lima, Joubert de Castro; Fortes, Reinaldo Silva; Ciferri, Ricardo Rodrigues; Silva, Rodrigo Rocha
    A tecnologia OLAP tem se consolidado há 20 anos e recentemente foi redesenhada para que suas dimensões, hierarquias e medidas possam suportar as particularidades dos dados textuais. A tarefa de organizar dados textuais de forma hierárquica pode ser resolvida com a construção de hierarquias de tópicos. Atualmente, a hierarquia de tópicos é definida apenas uma vez no cubo de dados, ou seja, para todo o \textit{lattice} de cuboides. No entanto, tal hierarquia é sensível ao conteúdo da coleção de documentos, portanto em um mesmo cubo de dados podem existir células com conteúdos completamente diferentes, agregando coleções de documentos distintas, provocando potenciais alterações na hierarquia de tópicos. Além disso, o segmento de texto utilizado na análise OLAP também influencia diretamente nos tópicos elencados por tal hierarquia. Neste trabalho, apresentamos um cubo de dados textual com múltiplas e dinâmicas hierarquias de tópicos. Múltiplas por serem construídas a partir de diferentes segmentos de texto e dinâmicas por serem construídas para cada célula do cubo. Outra contribuição deste trabalho refere-se à resposta das consultas multidimensionais. O estado da arte normalmente retorna os top-k documentos mais relevantes para um determinado tópico. Vamos além disso, retornando outros segmentos de texto, como os títulos mais significativos, resumos e parágrafos. A abordagem é projetada em quatro etapas adicionais, onde cada passo atenua um pouco mais o impacto da construção de várias hierarquias de tópicos e rankings de segmentos por célula de cubo. Experimentos que utilizam parte dos documentos da DBLP como uma coleção de documentos reforçam nossas hipóteses.
  • No Thumbnail Available
    Item
    PROLES : a customized process applied to software engineering laboratories and Small Business.
    (2017) Silva, Rodrigo Rocha; Kimura, Fernanda Yuri; Lima, Joubert de Castro; Bernardino, Jorge
    Most micro and small software development companies do not invest in improving the development-lifecycle processes. Given this, many researchers and software engineers have made proposals for software development processes with lower implementation costs and greater use of human and material resources. In general, it is required that work in academic laboratories be carried out in shorter deadlines, having groups of students practice techniques, such as different methodologies of requirements specification and fast responses to changes, ensuring the assertiveness in the value of deliveries of their projects. In this work, we customized the RUP and adopted Scrum practices, creating a new process called PROLES. PROLES was evaluated in a small company and also in a Software Engineering undergraduate course. PROLES has new particularities that differentiate it from the state of the art, such as a traceability matrix and a reference architecture. Such a matrix is obtained by the reference architecture, serving as a representation of the relationship among requirements and one or more system implementation components. The experimental results prove that PROLES can be adopted as an alternative to a systematic and less-intrusive process.
  • No Thumbnail Available
    Item
    qCube : efficient integration of range query operators over a high dimension data cube.
    (2013) Silva, Rodrigo Rocha; Lima, Joubert de Castro; Hirata, Celso Massaki
    Many decision support tasks involve range query operators such as Similar, Not Equal, Between, Greater or Less than and Some. Traditional cube approaches only use Equal operator in their summarized queries. Recent cube approaches implement range query operators, but they suffer from dimensionality problem, where a linear dimension increase consumes exponential storage space and runtime. Frag-Cubing and its extension, using bitmap index, are the most promising sequential solutions for high dimension data cubes, but they implement only Equal and Sub-cube query operators. In this paper, we implement a new high dimension sequential range cube approach, named Range Query Cube or just qCube. The qCube implements Equal, Not Equal, Distinct, Sub-cube, Greater or Less than, Some, Between, Similar and Top-k Similar query operators over a high dimension data cube. Comparative tests with qCube and Frag-Cubing use relations with 20, 30 or 60 dimensions, 5k distinct values on each dimension and 10 million tuples. In general, qCube has similar behavior when compared with Frag-Cubing, but it is faster to answer point and inquire queries. Frag-Cubing could not answer inquire queries with more than two Sub-cube operators in a relation with 30 dimensions, 5k cardinality and 10M tuples. In addition, qCube efficiently answered inquire queries from such a relation using six Sub-cube or Distinct operators. In general, complex queries with 30 operators, combining point, range and inquire operators, took less than 10 seconds to be answered byqCube. A massive qCube with 60 dimensions, 5k cardinality on each dimension and 100M tuples answered queries with five range operators, ten point operators and one inquire operator in less than 2 minutes.