arxivmath's Journal
[Most Recent Entries]
[Calendar View]
[Friends View]
Thursday, August 24th, 2006
| Time |
Event |
| 10:24p |
| | 10:24p |
Penalized Partial Least Squares Based on B-Splines Transformations. [math.ST/0608576]
Penalized Partial Least Squares Based on B-Splines Transformations. [math.ST/0608576]
Authors: Nicole Kraemer, Anne-Laure Boulesteix, Gerhard Tutz
We propose a novel method to model nonlinear regression problems by adapting
the principle of penalization to Partial Least Squares (PLS). Starting with a
generalized additive model, we expand the additive component of each variable
in terms of a generous amount of B-Splines basis functions. In order to prevent
overfitting and to obtain smooth functions, we estimate the regression model by
applying a penalized version of PLS. Although our motivation for penalized PLS
stems from its use for B-Splines transformed data, the proposed approach is
very general and can be applied to other penalty terms or to other dimension
reduction techniques. It turns out that penalized PLS can be computed virtually
as fast as PLS. We prove a close connection of penalized PLS to the solutions
of preconditioned linear systems. In the case of high-dimensional data, the new
method is shown to be an attractive competitor to other techniques for
estimating generalized additive models. If the number of predictor variables is
high compared to the number of examples, traditional techniques often suffer
from overfitting. We illustrate that penalized PLS performs well in these
situations.
read more at math updates on arXiv.org | | 10:24p |
Some new algebras of functions on topological groups arising from $G$-spaces. [math.DS/0608575]
Some new algebras of functions on topological groups arising from $G$-spaces. [math.DS/0608575]
Authors: Eli Glasner, Michael Megrelishvili
For a topological group G we introduce the algebra SUC(G) of \emph{strongly
uniformly continuous} functions. We show that SUC(G) contains the algebra
WAP(G) of weakly almost periodic functions as well as the algebras LE(G) and
Asp(G) of locally equicontinuous and Asplund functions respectively. For the
Polish groups of order preserving homeomorphisms of the unit interval and of
isometries of the Urysohn space of diameter 1, we show that SUC(G) is trivial.
We introduce the notion of fixed point on a class P of flows (P-fpp) and study
in particular groups with the SUC-fpp. We study the Roelcke algebra (= UC(G) =
right and left uniformly continuous functions) and SUC compactifications of the
groups S(N), of permutations of a countable set, and H(C), the group of
homeomorphisms of the Cantor set. For the first group we show that
WAP(G)=SUC(G)=UC(G) and also provide a concrete description of the
corresponding metrizable (in fact Cantor) semitopological semigroup
compactification. For the second group, in contrast, we show that SUC(G) is
properly contained in UC(G). We then deduce that for this group UC(G) does not
yield a right topological semigroup compactification.
read more at math updates on arXiv.org | | 10:24p |
| | 10:24p |
| | 10:24p |
| | 10:24p |
Intensional Models for the Theory of Types. [math.LO/0608571]
Intensional Models for the Theory of Types. [math.LO/0608571]
Authors: Reinhard Muskens
In this paper we define intensional models for the classical theory of types,
thus arriving at an intensional type logic ITL. Intensional models generalize
Henkin's general models and have a natural definition. As a class they do not
validate the axiom of Extensionality. We give a cut-free sequent calculus for
type theory and show completeness of this calculus with respect to the class of
intensional models via a model existence theorem. After this we turn our
attention to applications. Firstly, it is argued that, since ITL is truly
intensional, it can be used to model ascriptions of propositional attitude
without predicting logical omniscience. In order to illustrate this a small
fragment of English is defined and provided with an ITL semantics. Secondly, it
is shown that ITL models contain certain objects that can be identified with
possible worlds. Essential elements of modal logic become available within
classical type theory once the axiom of Extensionality is given up.
read more at math updates on arXiv.org | | 10:24p |
| | 10:24p |
Stable principal bundles and reduction of structure group. [math.AG/0608569]
Stable principal bundles and reduction of structure group. [math.AG/0608569]
Authors: Indranil Biswas
Let $E_G$ be a stable principal $G$--bundle over a compact connected Kaehler
manifold, where $G$ is a connected reductive linear algebraic group defined
over the complex numbers. Let $H\subset G$ be a complex reductive subgroup
which is not necessarily connected, and let $E_H\subset E_G$ be a holomorphic
reduction of structure group. We prove that $E_H$ is preserved by the
Einstein-Hermitian connection on $E_G$. Using this we show that if $E_H$ is a
minimal reductive reduction in the sense that there is no complex reductive
proper subgroup of $H$ to which $E_H$ admits a holomorphic reduction of
structure group, then $E_H$ is unique in the following sense: For any other
minimal reductive reduction $(H', E_{H'})$ of $E_G$, there is some element $g$
of $G$ such that $H'= g^{-1}Hg$ and $E_{H'}= E_Hg$. As an application, we give
an affirmative answer to a question of Balaji and Koll\'ar.
read more at math updates on arXiv.org | | 10:24p |
| | 10:24p |
Well-Balanced Schemes for the Initial Boundary Value Problem for 1D Scaler Conservation Laws. [math.
Well-Balanced Schemes for the Initial Boundary Value Problem for 1D Scaler Conservation Laws. [math.NA/0608567]
Authors: M. Nolte, D. Kroener
We consider well-balanced schemes for the following 1D scalar conservation
law with source term: d_t u + d_x f(u) + z'(x) b(u) = 0. More precisely, we are
interested in the numerical approximation of the initial boundary value problem
for this equation. While our main concern is a convergence result, we also have
to extend Otto's notion of entropy solutions to conservation laws with a source
term. To obtain uniqueness, we show that a generalization, the so-called
entropy process solution, is unique and coincides with the entropy solution. If
the initial and boundary data are essentially bounded, we can establish
convergence to the entropy solution. Showing that the numerical solutions are
bounded we can extract a weak*-convergent subsequence. Identifying its limit as
an entropy process solution requires some effort as we cannot use Kruzkov-type
entropy pairs here. We restrict ourselves to the Engquist-Osher flux and
identify the numerical entropy flux for an arbitrary entropy pair. By the
uniqueness result, the scheme then approximates the entropy solution and a
result by Vovelle then guarantees that the convergence is strong in L^p for
finite p.
read more at math updates on arXiv.org | | 10:24p |
| | 10:24p |
| | 10:24p |
| | 10:25p |
| | 10:25p |
| | 10:25p |
Propagation Time in Stochastic Communication Networks. [math.PR/0608561]
Propagation Time in Stochastic Communication Networks. [math.PR/0608561]
Authors: Jonathan Rowe, Boris Mitavskiy
Dynamical processes taking place on networks have received much attention in
recent years, especially on various models of random graphs (including small
world and scale free networks). They model a variety of phenomena, including
the spread of information on the Internet; the outbreak of epidemics in a
spatially structured population; and communication between randomly dispersed
processors in an ad hoc wireless network. Typically, research has concentrated
on the existence and size of a large connected component (representing, say,
the size of the epidemic) in a percolation model, or uses differential
equations to study the dynamics using a mean-field approximation in an infinite
graph. Here we investigate the time taken for information to propagate from a
single source through a finite network, as a function of the number of nodes
and the network topology. We assume that time is discrete, and that nodes
attempt to transmit to their neighbors in parallel, with a given probability of
success. We solve this problem exactly for several specific topologies, and use
a large-deviation theorem to derive general asymptotic bounds, which apply to
any family of networks where the diameter grows at least logarithmically in the
number of nodes. We use these bounds, for example, to show that a scale-free
network has propagation time logarithmic in the number of nodes, and inversely
proportional to the transmission probability.
read more at math updates on arXiv.org | | 10:25p |
| | 10:25p |
| | 10:25p |
| | 10:25p |
|
|