Person: Susi García, María Del Rosario
Loading...
First Name
María Del Rosario
Last Name
Susi García
Affiliation
Universidad Complutense de Madrid
Faculty / Institute
Estudios estadísticos
Department
Estadística y Ciencia de los Datos
Area
Estadística e Investigación Operativa
Identifiers
5 results
Search Results
Now showing 1 - 5 of 5
Item Inaccurate parameters in Gaussian Bayesian networks(2008) Gómez Villegas, Miguel Á.; Main Yaque, Paloma; Susi García, María Del RosarioTo determine the effect of a set of inaccurate parameters in Gaussian Bayesian networks, it is necessary to study the sensitive of the model. With this aim we propose a sensitivity analysis based on comparing two differents models: the original model with the initial parameters assigned to the Gaussian Bayesian network and the perturbed model obtain afther perturbing a ser of inaccurate parameters with specific characteristics.Item Análisis de sensibilidad en redes Bayesianas Gaussianas(2008) Susi García, María Del Rosario; Gómez Villegas, Miguel Angel; Maín Yaque, PalomaAl construir una Red Bayesiana se requiere que los expertos en el campo de aplicación especifiquen las dependencias entre las variables del problema e indiquen los parámetros que describen la red. Mediante este proceso de diseño y definición frecuentemente pueden asignarse erróneamente algunos parámetros y obtener resultados inadecuados. Por ello, surge la necesidad de introducir una medida de sensibilidad y robustez para Redes Bayesianas. En la bibliografía existente se han desarrollado diversas técnicas con este objetivo aunque los análisis disponibles estudian redes discretas o pequeños cambios alrededor de los parámetros. En esta memoria se desarrolla una metodología basada en una medida de discrepancia entre dos modelos de probabilidad; el inicial y el perturbado. La medida está basada en la seudodistancia de Kullback-Leibler y mediante ella, se compara la salida de la red para los dos modelos, con el fin de estudiar la sensibilidad o la robustez de una Red Bayesiana Gaussiana cuando se producen distintos cambios, pequeños o grandes, en los parámetros inciertos que describen la red. La metodología propuesta se concreta en el desarrollo de tres tipos de análisis para Redes Bayesianas Gaussianas, lo que en la memoria se denominan: análisis de sensibilidad de una vía, análisis de sensibilidad de n vías y análisis de robustez. Con los resultados obtenidos es posible determinar el comportamiento de una Red Bayesiana Gaussiana frente a todos los tipos de perturbaciones asociadas a los parámetros inciertos de la red.Item Perturbing the structure in Gaussian Bayesian networks(2009) Susi García, María Del Rosario; Navarro, H.; Main Yaque, Paloma; Gómez Villegas, Miguel Á.This paper introduces a n-way sensitivity analysis for Gaussian Bayesian networks where it studies the joint effect of variations in a set of similar parameters. The aim is to determine the sensitivity of the model when the parameters that describe the quantitative part are given by the structure of the graph. Therefore, with this analysis it studies the effect of uncertainty about the regression coefficients and the conditional variances of variables with their parents given in the graph.Item Extreme Inaccuracies In Gaussian Bayesian Networks(Journal Of Multivariate Analysis, 2008) Gómez Villegas, Miguel Ángel; Main Yaque, Paloma; Susi García, María Del RosarioTo evaluate the impact of model inaccuracies over the network’s output, after the evidence propagation, in a Gaussian Bayesian network, a sensitivity measure is introduced. This sensitivity measure is the Kullback–Leibler divergence and yields different expressions depending on the type of parameter to be perturbed, i.e. on the inaccurate parameter. In this work, the behavior of this sensitivity measure is studied when model inaccuracies are extreme,i.e. when extreme perturbations of the parameters can exist. Moreover, the sensitivity measure is evaluated for extreme situations of dependence between the main variables of the network and its behavior with extreme inaccuracies. This analysis is performed to find the effect of extreme uncertainty about the initial parameters of the model in a Gaussian Bayesian network and about extreme values of evidence. These ideas and procedures are illustrated with an example.Item Sensitivity Analysis in Gaussian Bayesian Networks Using a Divergence Measure(Communications in statistics. Theory and methods, 2007) Gómez Villegas, Miguel Ángel; Main Yaque, Paloma; Susi García, María Del RosarioThis article develops a method for computing the sensitivity analysis in a Gaussian Bayesian network. The measure presented is based on the Kullback–Leibler divergence and is useful to evaluate the impact of prior changes over the posterior marginal density of the target variable in the network. We find that some changes do not disturb the posterior marginal density of interest. Finally, we describe a method to compare different sensitivity measures obtained depending on where the inaccuracy was. An example is used to illustrate the concepts and methods presented.