Search results for "Residual sum of squares"
showing 6 items of 6 documents
A graphical model selection tool for mixed models
2017
Model selection can be defined as the task of estimating the performance of different models in order to choose the most parsimonious one, among a potentially very large set of candidate statistical models. We propose a graphical representation to be considered as an extension to the class of mixed models of the deviance plot proposed in the literature within the framework of classical and generalized linear models. This graphical representation allows, once a reduced number of models have been selected, to identify important covariates focusing only on the fixed effects component, assuming the random part properly specified. Nevertheless, we suggest also a standalone figure representing th…
Model interpretation from the additive elements of the PWRSS in GLMMs
2013
Generalized Linear Mixed models(GLMMs)have rapidly become a widely used tool for modelling clustered and longitudinal data with non-Normal responses. Although a large amount of work has been done in the literature on likelihood-based inference on GLMMs,little seems to have been done on the decomposition of the total variability associated to the different components of a mixed model.In this work we try to generalize the idea of likelihood additive elements Whittaker,1984), proposed in the context of GLMs,to the case of GLMMs by using the Penalized Weighted Residual Sum of Squares(PWRSS). The proposal is illustrated by means of areal application.
Enchytraeid population dynamics: Resource limitation and size-dependent mortality
2009
Abstract Enchytraeids are regarded as keystone soil organisms in forest ecosystems. Their abundance and biomass fluctuate widely. Predicting the consequences of anthropogenic disturbances requires an understanding of the mechanisms underlying enchytraeid population dynamics. Here I develop a simple model, which predicts that the type of dynamics is controlled by resource input rate. If fungal resource input is a discrete event once a year, an exponential growth phase is followed by starvation and sharp decline of enchytraeid abundance. Model simulations with three different forcing functions were compared to field data. Initial parameter values were obtained from various independent sources…
Variable selection in mixed models: a graphical approach
2014
Model selection can be defined as the task of estimating the performance of dif- ferent models in order to choose the (approximate) best one. The purpose of this article is to introduce an extension of the graphical representation of deviance proposed in the framework of classical and generalized linear models to the wider class of mixed models. The proposed plot is useful in determining which are the important explanatory variables conditioning on the random effects part. The applicability and the easy interpretation of the graph are illus- trated with a real data examples.
Algorithms for rational discrete least squares approximation
1975
In this paper an algorithm for the computation of a locally optimal polefree solution to the discrete rational least squares problem under a mild regularity condition is presented. It is based on an adaptation of projection methods [8], [12], [13], [14], [18], [19] to the modified Gaus-Newton method [4], [10]. A special device makes possible the direct handling of the infinitely many linear constraints present in this problem.
Prediction Model Selection and Spare Parts Ordering Policy for Efficient Support of Maintenance and Repair of Equipment
2010
The prediction model selection problem via variable subset selection is one of the most pervasive model selection problems in statistical applications. Often referred to as the problem of subset selection, it arises when one wants to model the relationship between a variable of interest and a subset of potential explanatory variables or predictors, but there is uncertainty about which subset to use. Several papers have dealt with various aspects of the problem but it appears that the typical regression user has not benefited appreciably. One reason for the lack of resolution of the problem is the fact that it has not been well defined. Indeed, it is apparent that there is not a single probl…