228 research outputs found
GAMLSS for high-dimensional data – a flexible approach based on boosting
Generalized additive models for location, scale and shape (GAMLSS) are a popular semi-parametric modelling approach that, in contrast to conventional GAMs, regress not only the expected mean but every distribution parameter (e.g. location, scale and shape) to a set of covariates. Current fitting procedures for GAMLSS are infeasible for high-dimensional data setups and require variable selection based on (potentially problematic) information criteria. The present work describes a boosting algorithm for high-dimensional GAMLSS that was developed to overcome these limitations. Specifically, the new algorithm was designed to allow the simultaneous estimation of predictor effects and variable selection. The proposed algorithm was applied to data of the Munich Rental Guide, which is used by
landlords and tenants as a reference for the average rent of a flat depending on its characteristics and spatial features. The net-rent predictions that resulted from the high-dimensional GAMLSS were found to be highly competitive while covariate-specific prediction intervals showed a major improvement over classical GAMs
E. Littmann & M. Hofner, Wörterbuch der Tigrē-Sprache, I
Caquot André. E. Littmann & M. Hofner, Wörterbuch der Tigrē-Sprache, I. In: Annales d'Ethiopie. Volume 2, année 1957. p. 266
Detection of the thermal radio continuum emission from the G9.62+0.19-F Hot Core
We present new high resolution and high sensitivity multi-frequency VLA radio continuum observations of the G9.62+0.19-F hot molecular core. We detect for the first time faint centimetric radio continuum emission at the position of the core. The centimetric continuum spectrum of the source is consistent with thermal emission from ionised gas. This is the first direct evidence that a newly born massive star is powering the G9.62+0.19-F hot core
Ritual Dynamics and Religious Change in the Roman Empire. Proceedings of the Eighth Workshop of the International Network Impact of Empire (Heidelberg, July 5-7, 2007)
Item does not contain fulltext377 p
Preface
Contains fulltext :
104468.pdf (Publisher’s version ) (Open Access)Eighth Workshop of the International Network Impact of Empire, 05 juli 2007377 p
Transreligiöse Integration als Faktor von Staatsbildung? Akbar und sein Gottesglaube in der europäischen Wahrnehmung
Flüchter A. Transreligiöse Integration als Faktor von Staatsbildung? Akbar und sein Gottesglaube in der europäischen Wahrnehmung. In: Eich P, Schmidt-Hofner S, Wieland C, eds. Der wiederkehrende Leviathan. Staatlichkeit und Staatswerdung in Spätantike und Früher Neuzeit. Akademie-Konferenzen. Vol 4. Heidelberg: Universitätsverlag Winter; 2011: 283-313
Variable Selection and Model Choice in Structured Survival Models
In many situations, medical applications ask for flexible survival models that allow to extend the classical Cox-model via the
inclusion of time-varying and nonparametric effects. These structured survival models are very flexible but additional
difficulties arise when model choice and variable selection is desired. In particular, it has to be decided which covariates
should be assigned time-varying effects or whether parametric modeling is sufficient for a given covariate. Component-wise
boosting provides a means of likelihood-based model fitting that enables simultaneous variable selection and model choice. We
introduce a component-wise likelihood-based boosting algorithm for survival data that permits the inclusion of both parametric
and nonparametric time-varying effects as well as nonparametric effects of continuous covariates utilizing penalized splines as
the main modeling technique. Its properties
and performance are investigated in simulation studies.
The new modeling approach is used to build a flexible survival model for
intensive care patients suffering from severe sepsis.
A software implementation is available to the interested reader
The Density Structure of Highly Compact H II Regions
We report the density structure of the ultracompact (UC) H II regions G35.20-1.74, G9.62+0.19-E, and G75.78+0.34-H2O. The density profiles are derived from radio continuum emission at wavelengths from 6 to 0.3 cm. In the case of G35.20-1.74, a cometary UC H II region with a core and a tail, the spectrum of the core varies as Sν~ν0.6, implying that the density structure is ne~r-2. The emission from the tail has a flatter spectrum, indicating that the density gradient is also negative but shallower. For the case of G9.62+0.19, which is an H II region complex with several components, the spectrum of the region designated component E is Sν~ν0.95, corresponding to ne~r-2.5. The steepest spectral index, Sν~ν1.4, is for the super UC H II region G75.78+0.34-H2O its density stratification may be as steep as ne~r-4. The actual density gradient may be smaller, owing to an exponential (rather than a power-law) density distribution or to the effects of finite spatial extent. The contribution from dust emission and some of the possible implications of these density distributions are briefly discussed
A study of the Keplerian accretion disk and precessing outflow in the massive protostar IRAS 20126+4104
We report on interferometric observations at 3.2 and 1.3 mm of the massive young stellar object IRAS 20126+4104 obtained in the C34S and CH3OH lines and in the continuum emission. The C34S data confirm the existence of a Keplerian disk, as already suggested by various authors. However, the mass of the central object is ~7 M_⊙, significantly less than previous estimates. We believe that such a discrepancy is due to the fact that the rotation curve is affected not only by the star but also by the mass in the innermost regions of the disk itself: this leads to an overestimate of the stellar mass when low-density tracers are used to study the velocity field over regions larger than a few seconds of arc (i.e. a few 0.01 pc). On the basis of the line profiles we speculate that accretion onto the star might be still occurring through the disk. This seems consistent with current models of high-mass star formation which predict an accretion luminosity equal to that of IRAS 20126+4104 for a 7 M_⊙ protostar. The CH3OH lines trace both the disk and the bipolar outflow previously detected in other molecules such as HCO^+, SiO, and H2. New H2 images obtained at 2.2 μm confirm that the outflow axis is undergoing precession. We elaborate a simple model that suitably fits the data thus allowing derivation of a few basic parameters of the precession
Model-based Boosting in R: A Hands-on Tutorial Using the R Package mboost
We provide a detailed hands-on tutorial for the R add-on package mboost. The package implements boosting for optimizing general risk functions utilizing component-wise (penalized) least squares estimates as base-learners for fitting various kinds of generalized linear and generalized additive models to potentially high-dimensional data. We give a theoretical background and demonstrate how mboost can be used to fit interpretable models of different complexity. As an example we use mboost to predict the body fat based on anthropometric measurements throughout the tutorial
- …
