Biomedical Engineering Technical Reports
http://hdl.handle.net/1853/25824
Papers and articles produced by Dr. Brani Vidakovic's research Mon, 20 Mar 2023 18:33:09 GMT2023-03-20T18:33:09ZUncertainty Analysis in Using Markov Chain Model to Predict Roof Life Cycle Performance
http://hdl.handle.net/1853/26261
Uncertainty Analysis in Using Markov Chain Model to Predict Roof Life Cycle Performance
Zhang, Yan; Vidakovic, Brani; Augenbroe, Godfried
Making decisions on building maintenance policies is an important topic in facility management. To evaluate different maintenance policies and make rational selection, both performance and maintenance cost of building components need to be of concern. For roofing sytem Markov Chain model has been developed to simulate the stochastic degrading process to evaluate the life cycle perfornance and cost. [Van Winden and Dekker 1998; Lounis et al. 1999] Taking value in a discrete state space, this model is especially appropriate when scaled rating regular inspections and related mainteance policies are implemented in large organizations. [Van Winden and Dekker 1998] However, many parameters in this Markov Chain model are associated with variance of significant magnitude. The propagation of these variances through the model will result in uncertainties in predicted life cycle performance and cost results. Without a solid uncertainty analysis on the simulation, decisions based on these simulation results can be unrealiable. In this paper we provide methods to estimate the range of parameter values and represent them in a probabilistic framwork. Monte Carlo method is used to analyze simulation output (life cycle cost and performance) variance propagated from these parameters through the model. These probablisitc informnation can be used to make better informed decisions. An example is provided to illustrate the Markov Chain model development, parameter identification method, Monte-Carlo uncertainty assessment and decision making with probabilistic information. It is shown that the uncertainty propagating through this process is not negligible and may significantly influence or even change the final decision
Presented at the 10DBMC International Conférence On Durability of Building Materials and Components, Lyon, France, 17-20 April 2005
Sat, 01 Jan 2005 00:00:00 GMThttp://hdl.handle.net/1853/262612005-01-01T00:00:00ZZhang, YanVidakovic, BraniAugenbroe, GodfriedMaking decisions on building maintenance policies is an important topic in facility management. To evaluate different maintenance policies and make rational selection, both performance and maintenance cost of building components need to be of concern. For roofing sytem Markov Chain model has been developed to simulate the stochastic degrading process to evaluate the life cycle perfornance and cost. [Van Winden and Dekker 1998; Lounis et al. 1999] Taking value in a discrete state space, this model is especially appropriate when scaled rating regular inspections and related mainteance policies are implemented in large organizations. [Van Winden and Dekker 1998] However, many parameters in this Markov Chain model are associated with variance of significant magnitude. The propagation of these variances through the model will result in uncertainties in predicted life cycle performance and cost results. Without a solid uncertainty analysis on the simulation, decisions based on these simulation results can be unrealiable. In this paper we provide methods to estimate the range of parameter values and represent them in a probabilistic framwork. Monte Carlo method is used to analyze simulation output (life cycle cost and performance) variance propagated from these parameters through the model. These probablisitc informnation can be used to make better informed decisions. An example is provided to illustrate the Markov Chain model development, parameter identification method, Monte-Carlo uncertainty assessment and decision making with probabilistic information. It is shown that the uncertainty propagating through this process is not negligible and may significantly influence or even change the final decisionWavelet Functional ANOVA, Bayesian False Discovery Rate, and Longitudinal Measurements of Oxygen Pressure in Rats
http://hdl.handle.net/1853/25943
Wavelet Functional ANOVA, Bayesian False Discovery Rate, and Longitudinal Measurements of Oxygen Pressure in Rats
Rosner, Gary L.; Vidakovic, Brani
Linear models can be functional in terms of independent or response variables or both. In functional ANOVA-type models often used to model longitudinal measurements and general time series, however, all components have a functional form. One of the main problems in inference using such models is the intrinsic dependence in “time” that makes pointwise inference difficult. We propose performing the inference in the wavelet domain instead of the time domain. Transformations by orthogonal wavelets preserve the structure of the linear model and, at the same time, decorrelate the data. The proposed methodology is applied to longitudinal measurements from experiments measuring oxygen pressure in tumor-bearing rats.
Sat, 01 Jan 2000 00:00:00 GMThttp://hdl.handle.net/1853/259432000-01-01T00:00:00ZRosner, Gary L.Vidakovic, BraniLinear models can be functional in terms of independent or response variables or both. In functional ANOVA-type models often used to model longitudinal measurements and general time series, however, all components have a functional form. One of the main problems in inference using such models is the intrinsic dependence in “time” that makes pointwise inference difficult. We propose performing the inference in the wavelet domain instead of the time domain. Transformations by orthogonal wavelets preserve the structure of the linear model and, at the same time, decorrelate the data. The proposed methodology is applied to longitudinal measurements from experiments measuring oxygen pressure in tumor-bearing rats.Discussion on Antoniadis and Fan "Regularization of Wavelet Approximations"
http://hdl.handle.net/1853/25941
Discussion on Antoniadis and Fan "Regularization of Wavelet Approximations"
Vidakovic, Brani
Anestis Antoniadis and Janqing Fan deserve congratulations for a wonderful and illuminating paper. Links among wavelet-based penalized function estimation, model selection, and now actively explored wavelet-shrinkage estimation, are intriguing and attracted attention of many researchers. Antoniadis and Fan provide numerous references. The nonlinear estimators resulting as optimal in the process of regularization, for some specific penalty functions, turn out to be the familiar hard- or soft-thresholding rules, or some of their sensible modifications. Simply speaking, the penalty function determines the estimation rule, and in many cases, a practicable and ad-hoc shrinkage rule can be linked to a regularization process under a reasonable penalty function. The authors explore the nature of penalty functions resulting in thresholding-type rules. They also show, that for a large class of penalty functions, corresponding shrinkage estimators are adaptively minimax and have other good sampling properties. My discussion will be directed toward the link of the regularization problem and Bayesian modeling and inference in the wavelet domain, which is only hinted by Antoniadis and Fan.
Accepted for publication in Journal of the American Statistical Association. The definitive version is available at http://www.amstat.org/publications/jasa/
Sat, 01 Sep 2001 00:00:00 GMThttp://hdl.handle.net/1853/259412001-09-01T00:00:00ZVidakovic, BraniAnestis Antoniadis and Janqing Fan deserve congratulations for a wonderful and illuminating paper. Links among wavelet-based penalized function estimation, model selection, and now actively explored wavelet-shrinkage estimation, are intriguing and attracted attention of many researchers. Antoniadis and Fan provide numerous references. The nonlinear estimators resulting as optimal in the process of regularization, for some specific penalty functions, turn out to be the familiar hard- or soft-thresholding rules, or some of their sensible modifications. Simply speaking, the penalty function determines the estimation rule, and in many cases, a practicable and ad-hoc shrinkage rule can be linked to a regularization process under a reasonable penalty function. The authors explore the nature of penalty functions resulting in thresholding-type rules. They also show, that for a large class of penalty functions, corresponding shrinkage estimators are adaptively minimax and have other good sampling properties. My discussion will be directed toward the link of the regularization problem and Bayesian modeling and inference in the wavelet domain, which is only hinted by Antoniadis and Fan.Denoising Ozone Concentration Measurements with BAMS Filtering
http://hdl.handle.net/1853/25940
Denoising Ozone Concentration Measurements with BAMS Filtering
Katul, Gabriel G.; Ruggeri, Fabrizio; Vidakovic, Brani
We propose a method for filtering self-similar geophysical signals corrupted by an antoregressive noise using a combination of non-decimated wavelet transform and a Bayesian model. In the application part, we consider separating the instrumentation noise from high frequency ozone concentration measurements sampled in the atmospheric boundary layer. The elicitation of priors needed to specify the statistical model in this application is guided by the well-known Kolmogorov K41-theory, which describes the statistical structure of the high frequency scalar concentration fluctuations.
Mon, 01 Jan 2001 00:00:00 GMThttp://hdl.handle.net/1853/259402001-01-01T00:00:00ZKatul, Gabriel G.Ruggeri, FabrizioVidakovic, BraniWe propose a method for filtering self-similar geophysical signals corrupted by an antoregressive noise using a combination of non-decimated wavelet transform and a Bayesian model. In the application part, we consider separating the instrumentation noise from high frequency ozone concentration measurements sampled in the atmospheric boundary layer. The elicitation of priors needed to specify the statistical model in this application is guided by the well-known Kolmogorov K41-theory, which describes the statistical structure of the high frequency scalar concentration fluctuations.