• Refine Query
  • Source
  • Publication year
  • to
  • Language
  • 414
  • 146
  • 56
  • 39
  • 37
  • 14
  • 9
  • 6
  • 6
  • 4
  • 4
  • 4
  • 4
  • 4
  • 4
  • Tagged with
  • 855
  • 111
  • 109
  • 92
  • 79
  • 78
  • 77
  • 76
  • 60
  • 60
  • 58
  • 53
  • 53
  • 47
  • 47
  • About
  • The Global ETD Search service is a free service for researchers to find electronic theses and dissertations. This service is provided by the Networked Digital Library of Theses and Dissertations.
    Our metadata is collected from universities around the world. If you manage a university/consortium/country archive and want to be added, details can be found on the NDLTD website.
51

DBM-tree: método de acesso métrico sensível à densidade local / DBM-tree: metric access method sensitive to local density data

Vieira, Marcos Rodrigues 28 May 2004 (has links)
Um espaço métrico é definido por um conjunto de objetos e uma função de distância métrica, que é utilizada para avaliar o nível de similaridade entre estes objetos. Isto permite a elaboração de Métodos de Acesso Métricos (MAMs) capazes de responder consultas por similaridade nesses conjuntos em um tempo reduzido. Em geral, esses MAMs são materializados através de uma estrutura hierárquica chamada de árvore métrica. Normalmente essas árvores são mantidas balanceadas, pois isto tende a manter a altura da árvore mínima, reduzindo o número de acessos a disco necessários para responder às consultas. No entanto, é difícil manter as estruturas balanceadas sem a existência de sobreposição entre os nós que cobrem regiões de alta densidade de objetos. O efeito disto é a degradação do tempo das consultas, pois várias subárvores devem ser analisadas para compor as consultas. Em outras palavras, minimizar a sobreposição entre os nós aumenta a eficiência das árvores métricas. Um meio efetivo para isto é flexibilizar o balanceamento das árvores métricas. Este trabalho apresenta um novo MAM dinâmico, chamado de DBM-tree (Density-Based Metric tree), que permite flexibilizar o balanceamento da estrutura, minimizando o grau de sobreposição entre os nós em regiões densas e, conseqüentemente, aumentando o seu desempenho para responder às consultas. Essa flexibilização é ajustada pelo usuário e é rigidamente controlada pela estrutura. A profundidade da árvore é maior em regiões de alta densidade, procurando um equilíbrio entre o número de acessos a disco para avaliar múltiplas subárvores e para a busca em profundidade em cada subárvore. A DBM-tree possui um algoritmo de otimização chamado de DBM-Slim-Down, que melhora o desempenho das árvores através da reorganização de elementos entre os seus nós. Os experimentos feitos com dados reais e sintéticos mostram que a DBM-tree supera em desempenho os MAMs tradicionais. Ela é, em média, 50% mais rápida que os MAMs tradicionais e reduz o número de acessos a disco e cálculos de distância em até 50%. Depois de executado o algoritmo DBM-Slim-Down, o seu desempenho melhorou em até 30% para as consultas por abrangência e aos vizinhos mais próximos. Ainda, a DBM-tree é escalável considerando tempo total de processamento, número de acessos a disco e de cálculos de distância em relação ao tamanho do conjunto de dados indexado. / A metric space is defined as a set of objects and a metric distance function that is used to measure the similarity between these objects. It allows the development of Metric Access Methods (MAMs) that are able to answer similarity queries in these datasets quickly. Usually these MAMs are materialized through a hierarchical structure called metric trees. These trees are kept balanced because it tends to maintain the height of the tree small, aiming to reduce the number of disk access required to answer queries. However, it is difficult to maintain the tree balanced without overlapping nodes covering a large number of objects, leading to the degradation of query performance. In other words, reducing the overlap among nodes increases the performance of metric trees. A possible solution is to relax the need to keep metric trees balanced. This work presents a new dynamic MAM called DBM-tree (Density-Based Metric tree), which changes the rule that imposes a rigid balancing policy, allowing a small amount of unbalancing in some regions of it. This unbalancing minimizes the degree of overlapping among some high-density nodes and, consequently, increases query answering performance. The amount of relaxation is set by the user and is strongly enforced in the tree. The height of the tree is higher in high-density regions, in order to keep a balance between searching in various subtrees and searching deeply in each subtree. The DBM-tree has an optimization algorithm called DBM-Slim-Down that improves the performance in trees through reorganizing the elements among its nodes. The experiments performed over synthetic and real-world datasets showed that the DBM-tree outperforms the traditional MAMs. The DBM-tree is, in average, 50% faster than traditional MAMs and reduces the number of distance calculations and disk accesses up to 50%. After executing the DBM-Slim-Down algorithm, the performance achieves improvements up to 30% for range and k-nearest neighbor queries. Moreover, the DBM-tree is scalable regarding time, number of disk accesses and distance calculations.
52

A Study Of The Metric Induced By The Robin Function

Borah, Diganta 07 1900 (has links) (PDF)
Let D be a smoothly bounded domain in Cn , n> 1. For each point p _ D, we have the Green function G(z, p) associated to the standard sum-of-squares Laplacian Δ with pole at p and the Robin constant __ Λ(p) = lim G(z, p) −|z − p−2n+2 z→p | at p. The function p _→ Λ(p) is called the Robin function for D. Levenberg and Yamaguchi had proved that if D is a C∞-smoothly bounded pseudoconvex domain, then the function log(−Λ) is a real analytic, strictly plurisubharmonic exhaustion function for D and thus induces a metric ds2 = n∂2 log(−Λ)(z) dzα ⊗ dzβ z ∂zα∂zβ α,β=1 on D, called the Λ-metric. For an arbitrary C∞-smoothly bounded domain, they computed the boundary asymptotics of Λ and its derivatives up to order 3, in terms of a defining function for the domain. As a consequence it was shown that the Λ-metric is complete on a C∞-smoothly bounded strongly pseudoconvex domain or a C∞-smoothly bounded convex domain. In this thesis, we study the boundary behaviour of the function Λ and its derivatives of all orders near a C2-smooth boundary point of an arbitrary domain. We compute the boundary asymptotics of the Λ-metric on a C∞-smoothly bounded pseudoconvex domain and as a consequence obtain that on a C∞-smoothly bounded strongly pseudoconvex domain, the Λ-metric is comparable to the Kobayashi metric (and hence to the Carath´eodory and the Bergman metrics). Using the boundary asymptotics of Λ and its derivatives, we calculate the holomorphic sectional curvature of the Λ-metric on a C∞-smoothly bounded strongly pseudoconvex domain at points on the inner normals and along the normal directions. The unit ball in Cn is also characterised among all C∞-smoothly bounded strongly convex domains on which the Λ-metric has constant negative holomorphic sectional curvature. Finally we study the stability of the Λ-metric under a C2 perturbation of a C∞-smoothly bounded pseudoconvex domain. (For equation pl refer the abstract pdf file)
53

DBM-tree: método de acesso métrico sensível à densidade local / DBM-tree: metric access method sensitive to local density data

Marcos Rodrigues Vieira 28 May 2004 (has links)
Um espaço métrico é definido por um conjunto de objetos e uma função de distância métrica, que é utilizada para avaliar o nível de similaridade entre estes objetos. Isto permite a elaboração de Métodos de Acesso Métricos (MAMs) capazes de responder consultas por similaridade nesses conjuntos em um tempo reduzido. Em geral, esses MAMs são materializados através de uma estrutura hierárquica chamada de árvore métrica. Normalmente essas árvores são mantidas balanceadas, pois isto tende a manter a altura da árvore mínima, reduzindo o número de acessos a disco necessários para responder às consultas. No entanto, é difícil manter as estruturas balanceadas sem a existência de sobreposição entre os nós que cobrem regiões de alta densidade de objetos. O efeito disto é a degradação do tempo das consultas, pois várias subárvores devem ser analisadas para compor as consultas. Em outras palavras, minimizar a sobreposição entre os nós aumenta a eficiência das árvores métricas. Um meio efetivo para isto é flexibilizar o balanceamento das árvores métricas. Este trabalho apresenta um novo MAM dinâmico, chamado de DBM-tree (Density-Based Metric tree), que permite flexibilizar o balanceamento da estrutura, minimizando o grau de sobreposição entre os nós em regiões densas e, conseqüentemente, aumentando o seu desempenho para responder às consultas. Essa flexibilização é ajustada pelo usuário e é rigidamente controlada pela estrutura. A profundidade da árvore é maior em regiões de alta densidade, procurando um equilíbrio entre o número de acessos a disco para avaliar múltiplas subárvores e para a busca em profundidade em cada subárvore. A DBM-tree possui um algoritmo de otimização chamado de DBM-Slim-Down, que melhora o desempenho das árvores através da reorganização de elementos entre os seus nós. Os experimentos feitos com dados reais e sintéticos mostram que a DBM-tree supera em desempenho os MAMs tradicionais. Ela é, em média, 50% mais rápida que os MAMs tradicionais e reduz o número de acessos a disco e cálculos de distância em até 50%. Depois de executado o algoritmo DBM-Slim-Down, o seu desempenho melhorou em até 30% para as consultas por abrangência e aos vizinhos mais próximos. Ainda, a DBM-tree é escalável considerando tempo total de processamento, número de acessos a disco e de cálculos de distância em relação ao tamanho do conjunto de dados indexado. / A metric space is defined as a set of objects and a metric distance function that is used to measure the similarity between these objects. It allows the development of Metric Access Methods (MAMs) that are able to answer similarity queries in these datasets quickly. Usually these MAMs are materialized through a hierarchical structure called metric trees. These trees are kept balanced because it tends to maintain the height of the tree small, aiming to reduce the number of disk access required to answer queries. However, it is difficult to maintain the tree balanced without overlapping nodes covering a large number of objects, leading to the degradation of query performance. In other words, reducing the overlap among nodes increases the performance of metric trees. A possible solution is to relax the need to keep metric trees balanced. This work presents a new dynamic MAM called DBM-tree (Density-Based Metric tree), which changes the rule that imposes a rigid balancing policy, allowing a small amount of unbalancing in some regions of it. This unbalancing minimizes the degree of overlapping among some high-density nodes and, consequently, increases query answering performance. The amount of relaxation is set by the user and is strongly enforced in the tree. The height of the tree is higher in high-density regions, in order to keep a balance between searching in various subtrees and searching deeply in each subtree. The DBM-tree has an optimization algorithm called DBM-Slim-Down that improves the performance in trees through reorganizing the elements among its nodes. The experiments performed over synthetic and real-world datasets showed that the DBM-tree outperforms the traditional MAMs. The DBM-tree is, in average, 50% faster than traditional MAMs and reduces the number of distance calculations and disk accesses up to 50%. After executing the DBM-Slim-Down algorithm, the performance achieves improvements up to 30% for range and k-nearest neighbor queries. Moreover, the DBM-tree is scalable regarding time, number of disk accesses and distance calculations.
54

Understanding Noise and Structure behind Metric Spaces

Wang, Dingkang 20 October 2021 (has links)
No description available.
55

A study of Monoidal t-norm based Logic

Toloane, Ellen Mohau 07 February 2014 (has links)
The logical system MTL (for Monoidal t-norm Logic) is a formalism of the logic of left-continuous t-norms, which are operations that arise in the study of fuzzy sets and fuzzy logic. The objective is to investigate the important results on MTL and collect them together in a coherent form. The main results considered will be the completeness results for the logic with respect to MTL-algebras, MTL-chains (linearly ordered MTL-algebras) and standard MTL-algebras (left-continuous t-norm algebras). Completeness of MTL with respect to standard MTL-algebras means that MTL is indeed the logic of left-continuous t-norms. The logical system BL (for Basic Logic) is an axiomatic extension of MTL; we will consider the same completeness results for BL; that is we will show that BL is complete with respect to BL-algebras, BL-chains and standard BL-algebras (continuous t-norm algebras). Completeness of BL with respect to standard BL-algebras means that BL is the logic of continuous t-norms.
56

Some effects of the compulsory use of metric weights and measures: a study of the results of the compulsory use of metric weights and measures in Brazil, Argentina, Uruguay, and Paraguay

Yorke, Gertrude Cushing January 1942 (has links)
Thesis (Ed.D.)--Boston University
57

Algorithms for Optimal Transport and Wasserstein Distances

Schrieber, Jörn 14 February 2019 (has links)
No description available.
58

Metric system conversion for hospital dietary departments

Mih, Mew Y January 2010 (has links)
Digitized by Kansas Correctional Industries
59

New Statistical Methods to Get the Fractal Dimension of Bright Galaxies Distribution from the Sloan Digital Sky Survey Data

Wu, Yongfeng January 2007 (has links) (PDF)
No description available.
60

Perceptually-based Comparison of Image Similarity Metrics

Russell, Richard, Sinha, Pawan 01 July 2001 (has links)
The image comparison operation ??sessing how well one image matches another ??rms a critical component of many image analysis systems and models of human visual processing. Two norms used commonly for this purpose are L1 and L2, which are specific instances of the Minkowski metric. However, there is often not a principled reason for selecting one norm over the other. One way to address this problem is by examining whether one metric better captures the perceptual notion of image similarity than the other. With this goal, we examined perceptual preferences for images retrieved on the basis of the L1 versus the L2 norm. These images were either small fragments without recognizable content, or larger patterns with recognizable content created via vector quantization. In both conditions the subjects showed a consistent preference for images matched using the L1 metric. These results suggest that, in the domain of natural images of the kind we have used, the L1 metric may better capture human notions of image similarity.

Page generated in 0.0469 seconds