The modern scientific process often involves the development of a predictive computational model. To improve its accuracy, a computational model can be calibrated to a set of experimental data. A variety of validation metrics can be used to quantify this process. Some of these metrics have direct physical interpretations and a history of use, while others, especially those for probabilistic data, are more difficult to interpret. In this work, a variety of validation metrics are used to quantify the accuracy of different calibration methods. Frequentist and Bayesian perspectives are used with both fixed effects and mixed-effects statistical models. Through a quantitative comparison of the resulting distributions, the most accurate calibration method can be selected. Two examples are included which compare the results of various validation metrics for different calibration methods. It is quantitatively shown that, in the presence of significant laboratory biases, a fixed effects calibration is significantly less accurate than a mixed-effects calibration. This is because the mixed-effects statistical model better characterizes the underlying parameter distributions than the fixed effects model. The results suggest that validation metrics can be used to select the most accurate calibration model for a particular empirical model with corresponding experimental data.
The thermal challenge problem has been developed at Sandia National Laboratories as a testbed for demonstrating various types of validation approaches and prediction methods. This report discusses one particular methodology to assess the validity of a computational model given experimental data. This methodology is based on Bayesian Belief Networks (BBNs) and can incorporate uncertainty in experimental measurements, in physical quantities, and model uncertainties. The approach uses the prior and posterior distributions of model output to compute a validation metric based on Bayesian hypothesis testing (a Bayes' factor). This report discusses various aspects of the BBN, specifically in the context of the thermal challenge problem. A BBN is developed for a given set of experimental data in a particular experimental configuration. The development of the BBN and the method for ''solving'' the BBN to develop the posterior distribution of model output through Monte Carlo Markov Chain sampling is discussed in detail. The use of the BBN to compute a Bayes' factor is demonstrated.
This document provides verification test results for normal, lognormal, and uniform distributions that are used in Sandia's Latin Hypercube Sampling (LHS) software. The purpose of this testing is to verify that the sample values being generated in LHS are distributed according to the desired distribution types. The testing of distribution correctness is done by examining summary statistics, graphical comparisons using quantile-quantile plots, and format statistical tests such as the Chisquare test, the Kolmogorov-Smirnov test, and the Anderson-Darling test. The overall results from the testing indicate that the generation of normal, lognormal, and uniform distributions in LHS is acceptable.
The ground truth program used simulations as test beds for social science research methods. The simulations had known ground truth and were capable of producing large amounts of data. This allowed research teams to run experiments and ask questions of these simulations similar to social scientists studying real-world systems, and enabled robust evaluation of their causal inference, prediction, and prescription capabilities. We tested three hypotheses about research effectiveness using data from the ground truth program, specifically looking at the influence of complexity, causal understanding, and data collection on performance. We found some evidence that system complexity and causal understanding influenced research performance, but no evidence that data availability contributed. The ground truth program may be the first robust coupling of simulation test beds with an experimental framework capable of teasing out factors that determine the success of social science research.
All disciplines that use models to predict the behavior of real-world systems need to determine the accuracy of the models’ results. Techniques for verification, validation, and uncertainty quantification (VVUQ) focus on improving the credibility of computational models and assessing their predictive capability. VVUQ emphasizes rigorous evaluation of models and how they are applied to improve understanding of model limitations and quantify the accuracy of model predictions.