No description available.

Thesis (M.S.)--University of South Africa. / Includes bibliographical references (leaves 200-202).

This thesis focuses on ridge regression (RR) and least absolute shrinkage and selection operator (lasso). Ridge properties are being investigated in great detail which include studying the bias, the variance and the mean squared error as a function of the tuning parameter. We also study the convexity of the trace of the mean squared error in terms of the tuning parameter. In addition, we examined some special properties of RR for factorial experiments. Not only do we review ridge properties, we also review lasso properties because they are somewhat similar. Rather than shrinking the estimates toward zero in RR, the lasso is able to provide a sparse solution, setting many coefficient estimates exaclty to zero. Furthermore, we try a new approach to solve the lasso problem by formulating it as a bilevel problem and implementing a new algorithm to solve this bilevel program. / Graduate

No description available.

No description available.

No description available.

[pt] Nas aplicações em computação gráfica e processamento de imagens, curvas e superfícies implícitas têm sido reconhecidas como a representação mais útil de objetos 2D ou 3D, principalmente porque elas permitem a descrição de formas complexas por uma fórmula. A maioria dos métodos implícitos usam curvas algébricas para aproximar globalmente a fronteira do objeto em uma imagem binária. Quando a forma do objeto é complexa, é comum elevar o grau da curva a fim de obter mais precisão na aproximação. Uma solução alternativa é decompor hierarquicamente o domínio em partes compactas e obter aproximações locais para o objeto em cada parte, e então juntar os pedaços com o objetivo de obter uma descrição global do objeto. O principal objetivo deste trabalho é apresentar um novo método de aproximação de curvas implícitas a partir de pontos esparsos que melhora o estado da arte / [en] In the field of computer vision and image analysis, implicit curves and surfaces have been recognized as the most useful representation for 2D or 3D objects, mainly because they allow description of shapes by a formula. Most of implicit methods uses algebraic curves to fit globally the frontier of the foreground in a binary image. When the foreground shape is complex, it is common to elevate the curve degree in order to obtain more precision on the approximation. An alternative solution is to decompose the domain hierarchicaly in compact parts and obtain local approximation for the object in each part, and then patch all together in order to obtain a global description of the object. The main objective of this work is to present a new method for implicit curve fitting from sparse point that improves the state of the art

Thesis (Ph. D.)--University of California, San Diego, 2007. / Title from first page of PDF file (viewed October 4, 2007). Available via ProQuest Digital Dissertations. Vita. Includes bibliographical references.

Thesis (M.S.)--Ohio University, November, 1996. / Title from PDF t.p.

Thesis (Ph. D.)--University of Washington, 1998. / Vita. Includes bibliographical references (p. [115]-122).