No description available.

No description available.

Thesis (Ph. D.)--University of Kentucky, 2004. / Title from document title page (viewed Jan. 7, 2005). Document formatted into pages; contains x, 180p. : ill. Includes abstract and vita. Includes bibliographical references (p. 173-178).

Bibliography: leaves 90-93.

In many emerging applications, such as sensor networks, location-based services, and data integration, the database is inherently uncertain. To handle a large amount of uncertain data, probabilistic databases have been recently proposed, where probabilistic queries are enabled to provide answers with statistical guarantees. In this thesis, we study the important issues of managing the quality of a probabilistic database. We first address the problem of measuring the ambiguity, or quality, of a probabilistic query. This is accomplished by computing the PWS-quality score, a recently proposed measure for quantifying the ambiguity of query answers under the possible world semantics. We study the computation of the PWS-quality for the top-k query. This problem is not trivial, since directly computing the top-k query score is computationally expensive. To tackle this challenge, we propose efficient approximate algorithms for deriving the quality score of a top-k query. We have performed experiments on both synthetic and real data to validate their performance and accuracy. Our second contribution is to study how to use the PWS-quality score to coordinate the process of cleaning uncertain data. Removing ambiguous data from a probabilistic database can often give us a higher-quality query result. However, this operation requires some external knowledge (e.g., an updated value from a sensor source), and is thus not without cost. It is important to choose the correct object to clean, in order to (1) achieve a high quality gain, and (2) incur a low cleaning cost. In this thesis, we examine different cleaning methods for a probabilistic top-k query. We also study an interesting problem where different query users have their own budgets available for cleaning. We demonstrate how an optimal solution, in terms of the lowest cleaning costs, can be achieved, for probabilistic range and maximum queries. An extensive evaluation reveals that these solutions are highly efficient and accurate. / published_or_final_version / Computer Science / Master / Master of Philosophy

Thesis (M. Sc.)--University of Adelaide, 1987. / Includes bibliographical references (leaves 90-93).

Em teoria dos números, o teorema de Erdös-Kac, também conhecido como o teorema fundamental de teoria probabilística dos números, diz que se w(n) denota a quantidade de fatores primos distintos de n, então a sequência de funções de distribuições N definidas por FN(x) = (1/N) #{n <= N : (w(n) log log N)/(log log N)^(1/2)} <= x}, converge uniformemente sobre R para a distribuição normal padrão. Neste trabalho desenvolvemos todos os teoremas necessários para uma demonstração analítica, que nos permitirá encontrar a ordem de erro da convergência acima. / In number theory, the Erdös-Kac theorem, also known as the fundamental theorem of probabilistic number theory, states that if w(n) is the number of distinct prime factors of n, then the sequence of distribution functions N, defined by FN(x) = (1/N) #{n <= N : (w(n) log log N)/(log log N)^(1/2)} <= x}, converges uniformly on R to the standard normal distribution. In this work we developed all theorems needed to an analytic demonstration, which will allow us to find an order of error of the above convergence.

Em teoria dos números, o teorema de Erdös-Kac, também conhecido como o teorema fundamental de teoria probabilística dos números, diz que se w(n) denota a quantidade de fatores primos distintos de n, então a sequência de funções de distribuições N definidas por FN(x) = (1/N) #{n <= N : (w(n) log log N)/(log log N)^(1/2)} <= x}, converge uniformemente sobre R para a distribuição normal padrão. Neste trabalho desenvolvemos todos os teoremas necessários para uma demonstração analítica, que nos permitirá encontrar a ordem de erro da convergência acima. / In number theory, the Erdös-Kac theorem, also known as the fundamental theorem of probabilistic number theory, states that if w(n) is the number of distinct prime factors of n, then the sequence of distribution functions N, defined by FN(x) = (1/N) #{n <= N : (w(n) log log N)/(log log N)^(1/2)} <= x}, converges uniformly on R to the standard normal distribution. In this work we developed all theorems needed to an analytic demonstration, which will allow us to find an order of error of the above convergence.

Dans le premier chapitre de cette thèse, nous passons en revue les outils de la théorie analytique des nombres qui seront utiles pour la suite. Nous faisons aussi un survol des entiers y−friables, c’est-à-dire des entiers dont chaque facteur premier est plus petit ou égal à y. Au deuxième chapitre, nous présenterons des problèmes classiques de la théorie des nombres probabiliste et donnerons un bref historique d’une classe de fonctions arithmétiques sur un espace probabilisé. Le problème de Erdos sur la table de multiplication demande quel est le nombre d’entiers distincts apparaissant dans la table de multiplication N × N. L’ordre de grandeur de cette quantité a été déterminé par Kevin Ford (2008). Dans le chapitre 3 de cette thèse, nous étudions le nombre d’ensembles y−friables de la table de multiplication N × N. Plus concrètement, nous nous concentrons sur le changement du comportement de la fonction A(x, y) par rapport au domaine de y, où A(x, y) est une fonction qui compte le nombre d’entiers y− friables distincts et inférieurs à x qui peuvent être représentés comme le produit de deux entiers y− friables inférieurs à p x. Dans le quatrième chapitre, nous prouvons un théorème de Erdos-Kac modifié pour l’ensemble des entiers y− friables. Si !(n) est le nombre de facteurs premiers distincts de n, nous prouvons que la distribution de !(n) est gaussienne pour un certain domaine de y en utilisant la méthode des moments. / The object of the first chapter of this thesis is to review the materials and tools in analytic number theory which are used in following chapters. We also give a survey on the development concerning the number of y−smooth integers, which are integers free of prime factors greater than y. In the second chapter, we shall give a brief history about a class of arithmetical functions on a probability space and we discuss on some well-known problems in probabilistic number theory. We present two results in analytic and probabilistic number theory. The Erdos multiplication table problem asks what is the number of distinct integers appearing in the N × N multiplication table. The order of magnitude of this quantity was determined by Kevin Ford (2008). In chapter 3 of this thesis, we study the number of y−smooth entries of the N × N multiplication. More concretely, we focus on the change of behaviour of the function A(x,y) in different ranges of y, where A(x,y) is a function that counts the number of distinct y−smooth integers less than x which can be represented as the product of two y−smooth integers less than p x. In Chapter 4, we prove an Erdos-Kac type of theorem for the set of y−smooth integers. If !(n) is the number of distinct prime factors of n, we prove that the distribution of !(n) is Gaussian for a certain range of y using method of moments.