The effect of block parameter perturbations in Gaussian Bayesian networks: Sensitivity and robustness

BackgroundA number of studies have previously demonstrated that “goodness of fit” is insufficient in reliably classifying the credibility of a biological model. Robustness and/or sensitivity analysis is commonly employed as a secondary method for evaluating the suitability of a particular model. The results of such analyses invariably depend on the particular parameter set tested, yet many parameter values for biological models are uncertain.ResultsHere, we propose a novel robustness analysis that aims to determine the “common robustness” of the model with multiple, biologically plausible parameter sets, rather than the local robustness for a particular parameter set. Our method is applied to two published models of the Arabidopsis circadian clock (the one-loop [1] and two-loop [2] models). The results reinforce current findings suggesting the greater reliability of the two-loop model and pinpoint the crucial role of TOC1 in the circadian network.ConclusionsConsistent Robustness Analysis can indicate both the relative plausibility of different models and also the critical components and processes controlling each model.

Embedding a large vocabulary speech recognition system in mobile devices requires a reduced acoustic model obtained by eliminating redundant model parameters. In conventional optimization methods based on the minimum description length (MDL) criterion, a binary Gaussian tree is built at each state of a hidden Markov model by iteratively finding and merging similar mixture components. An optimal subset of the tree nodes is then selected to generate a downsized acoustic model. To obtain a better binary Gaussian tree by improving the process of finding the most similar Gaussian components, this paper proposes a new distance measure that exploits the difference in likelihood values for cases before and after two components are combined. The mixture weight of Gaussian components is also introduced in the component merging step. Experimental results show that the proposed method outperforms MDL-based optimization using either a Kullback-Leibler (KL) divergence or weighted KL divergence measure. The proposed method could also reduce the acoustic model size by 50% with less than a 1.5% increase in error rate compared to a baseline system.

Ecosystem models are useful tools for understanding ecological processes and for sustainable management of resources. In biogeochemical field, numerical models have been widely used for investigating carbon dynamics under global changes from site to regional and global scales. However, it is still challenging to optimize parameters and estimate parameterization uncertainty for complex process‐based models such as the Erosion Deposition Carbon Model (EDCM), a modified version of CENTURY, that consider carbon, water, and nutrient cycles of ecosystems. This study was designed to conduct the parameter identifiability, optimization, sensitivity, and uncertainty analysis of EDCM using our developed EDCM‐Auto, which incorporated a comprehensive R package—Flexible Modeling Framework (FME) and the Shuffled Complex Evolution (SCE) algorithm. Using a forest flux tower site as a case study, we implemented a comprehensive modeling analysis involving nine parameters and four target variables (carbon and water fluxes) with their corresponding measurements based on the eddy covariance technique. The local sensitivity analysis shows that the plant production‐related parameters (e.g., PPDF1 and PRDX) are most sensitive to the model cost function. Both SCE and FME are comparable and performed well in deriving the optimal parameter set with satisfactory simulations of target variables. Global sensitivity and uncertainty analysis indicate that the parameter uncertainty and the resulting output uncertainty can be quantified, and that the magnitude of parameter‐uncertainty effects depends on variables and seasons. This study also demonstrates that using the cutting‐edge R functions such as FME can be feasible and attractive for conducting comprehensive parameter analysis for ecosystem modeling.

Development of theoretical approaches to the control of a water-treatment system

Control systems operate under the assumption that sensors are trustworthy. Yet, when communication channels are unprotected or sensors are accessible from networked stations, malicious users can compromise the system by spoofing the measured information. We consider a linear time-invariant system with a single sensor, where the state is estimated by a Kalman filter. We assume the presence of an attacker with the ability to modify the measurements arbitrarily, which are then processed by the Kalman filter for as long as the attacker remains undetected. The objective of the attacker is to maximize the mean square error of the Kalman filter. We adopt a notion of attack stealthiness based on the Kullback–Leibler divergence measure, and characterize the worst case degradation induced by an attacker with a fixed stealthiness level. Additionally, we characterize optimal attack strategies that achieve our bound of performance degradation, thereby proving tightness of our result.

<p>Complex network approaches have been recently emerging as novel and complementary concepts of nonlinear time series analysis which are able to unveil many features that are hidden to more traditional analysis methods. In this talk, we focus on one particular approach of ordinal pattern transition networks (OPTNs) for characterizing time series data. In particular, we introduce a suite of OPTN based complexity measures to infer the coupling direction between two dynamical systems from pairs of time series. For several examples of both coupled stochastic processes and chaotic Henon maps, we demonstrate that our approach is able to successfully identify interaction delays of both unidirectional and bidirectional coupling configurations.</p><p>Furthermore, we focus on applying these methods to characterize the recent extreme drought events in the semiarid region of Northeast Brazil (NEB) where has been experiencing a continuous dry condition since 2012. Therefore, we propose a three-step strategy to establish the episodic coupling directions on intraseasonal time scales from the surrounding ocean to the precipitation patterns in the NEB, focusing on the distinctive roles of the oceans during the recent extreme drought events of 2012-2013 and 2015-2016. Our algorithm involves: (i) computing drought period length from daily precipitation anomalies to capture extreme drought events, (ii) characterizing the episodic coupling delays from the surrounding oceans to the precipitation by applying Kullback-Leibler divergence (KLD) of complexity measure which is based on OPTN representation of time series, and (iii) calculating the ratio of high temperature in the ocean during the extreme drought events with proper time lags that are identified by KLD measures. From the viewpoint of climatology, our analysis provides data-based evidence of showing significant influence from the North Atlantic in 2012-2013 to the NEB, but in 2015-2016 the Pacific played a dominant role than that of the Atlantic. The episodic intra-seasonal time scale properties are potential for monitoring and forecasting droughts in the NEB, in order to propose strategies for drought impacts reduction.</p><p>In conclusion, our results suggest that ordinal partition transition networks can be used as complementary tools for causal inference tasks and provide insights into the potentials and theoretical foundations of time series networks.</p><p>

Purpose: In particle therapy, treatment planning and evaluation are frequently based on biological models to estimate the relative biological effectiveness (RBE) or the equivalent dose in 2 Gy fractions (EQD2). In the context of the linear-quadratic model, these quantities depend on biological parameters (α, β) for ions as well as for the reference radiation and on the dose per fraction. The needed biological parameters as well as their dependency on ion species and ion energy typically are subject to large (relative) uncertainties of up to 20–40% or even more. Therefore it is necessary to estimate the resulting uncertainties in e.g. RBE or EQD2 caused by the uncertainties of the relevant input parameters. Methods: We use a variance-based sensitivity analysis (SA) approach, in which uncertainties in input parameters are modeled by random number distributions. The evaluated function is executed 10{sup 4} to 10{sup 6} times, each run with a different set of input parameters, randomly varied according to their assigned distribution. The sensitivity S is a variance-based ranking (from S = 0, no impact, to S = 1, only influential part) of the impact of input uncertainties. The SA approach is implemented for carbon ion treatment plans on 3D patient data,more » providing information about variations (and their origin) in RBE and EQD2. Results: The quantification enables 3D sensitivity maps, showing dependencies of RBE and EQD2 on different input uncertainties. The high number of runs allows displaying the interplay between different input uncertainties. The SA identifies input parameter combinations which result in extreme deviations of the result and the input parameter for which an uncertainty reduction is the most rewarding. Conclusion: The presented variance-based SA provides advantageous properties in terms of visualization and quantification of (biological) uncertainties and their impact. The method is very flexible, model independent, and enables a broad assessment of uncertainties. Supported by DFG grant WI 3745/1-1 and DFG cluster of excellence: Munich-Centre for Advanced Photonics.« less

An interactive graphical interface tool for parameter calibration, sensitivity analysis, uncertainty analysis, and visualization for the Soil and Water Assessment Tool

In this work, we develop methodologies for analyzing and cross comparing metabolic models. We investigate three important metabolic networks to discuss the complexity of biological organization of organisms, modeling, and system properties. In particular, we analyze these metabolic networks because of their biotechnological and basic science importance: the photosynthetic carbon metabolism in a general leaf, the Rhodobacter spheroides bacterium, and the Chlamydomonas reinhardtii alga. We adopt single- and multi-objective optimization algorithms to maximize the CO 2 uptake rate and the production of metabolites of industrial interest or for ecological purposes. We focus both on the level of genes (e.g., finding genetic manipulations to increase the production of one or more metabolites) and on finding concentration enzymes for improving the CO 2 consumption. We find that R. spheroides is able to absorb an amount of CO 2 until 57.452 mmol h (-1) gDW (-1) , while C. reinhardtii obtains a maximum of 6.7331. We report that the Pareto front analysis proves extremely useful to compare different organisms, as well as providing the possibility to investigate them with the same framework. By using the sensitivity and robustness analysis, our framework identifies the most sensitive and fragile components of the biological systems we take into account, allowing us to compare their models. We adopt the identifiability analysis to detect functional relations among enzymes; we observe that RuBisCO, GAPDH, and FBPase belong to the same functional group, as suggested also by the sensitivity analysis.

Robustness analysis is a combination of reliability and geometrical strength analysis using a strain technique. It refers to the ability of a network to resist deformations caused by the largest undetectable blunders. The displacement of each point in the network is computed in order to measure the robustness of the network. This paper tries to optimally design a geodetic network in the sense of high reliability and geometrical strength. For this purpose, a metaheuristic method called the shuffled frog leaping algorithm (SFLA) is used to solve the first order design problem in which the geometric configuration of the network is optimised. Such algorithms have been developed to determine high quality solutions to complex optimisation problems. The efficiency of the method is demonstrated using a synthetic network example. The results show that the displacements can be decreased by maximising the minimum redundancy number in the network. This procedure can yield both reliable and robust networks.

Population viability analysis (PVA) is widely used to assess population-level impacts of environmental changes on species. When combined with sensitivity analysis, PVA yields insights into the effects of parameter and model structure uncertainty. This helps researchers prioritize efforts for further data collection so that model improvements are efficient and helps managers prioritize conservation and management actions. Usually, sensitivity is analyzed by varying one input parameter at a time and observing the influence that variation has over model outcomes. This approach does not account for interactions among parameters. Global sensitivity analysis (GSA) overcomes this limitation by varying several model inputs simultaneously. Then, regression techniques allow measuring the importance of input-parameter uncertainties.In many conservation applications, the goal of demographic modeling is to assess how different scenarios of impact or management cause changes in a population. This is challenging because the uncertainty of input-parameter values can be confounded with the effect of impacts and management actions. We developed a GSA method that separates model outcome uncertainty resulting from parameter uncertainty from that resulting from projected ecological impacts or simulated management actions, effectively separating the 2 main questions that sensitivity analysis asks. We applied this method to assess the effects of predicted sea-level rise on Snowy Plover (Charadrius nivosus). A relatively small number of replicate models (approximately 100) resulted in consistent measures of variable importance when not trying to separate the effects of ecological impacts from parameter uncertainty. However, many more replicate models (approximately 500) were required to separate these effects. These differences are important to consider when using demographic models to estimate ecological impacts of management actions.

Aerodynamic response of structures with parametric uncertainties

A probabilistic framework and the supporting tools are developed and demonstrated to investigate the out-of-plane properties of a composite lamina. The framework methodology consists of a sensitivity analysis to determine the most influential input parameters followed by a probabilistic analysis to quantify the effects of the uncertainty in these parameters on lamina properties. To demonstrate the framework, sensitivity analysis is performed through finite element analysis of unit cell models to identify the most influential constituent and configuration parameters on lamina behavior, specifically on out-of-plane lamina properties. The effects of uncertainty in these input parameters are then quantified through Monte Carlo simulation using probabilistic distributions to represent each input under consideration. One of three distribution types (Normal, Lognormal, and Weibull) can be assigned, and a different distribution type can represent each input parameter. To study configuration effects, a representative volume element (RVE) consisting of two different fiber materials is investigated. The location and percent composition of the two different fiber materials are varied to analyze their effects on out-of-plane lamina properties.

A review of sensitivity analysis methods in building energy analysis

Factor analysis is a statistical method used to describe a set of variables based on common dimensions. Factor analysis that is often used is the classical factor analysis with principal components method. Classical factor analysis can not work properly if the data contained many outliers. In order factor analysis remains optimal in explaining a set of variables even in conditions of data containing many outliers, we need a robust estimator. Through factor analysis is expected to obtain robust high accuracy analysis results for data containing many outliers. Estimator fast-MCD is one of the robust estimator that aims to get the smallest determinant of the covariance matrix. Robust factor analysis with fast-MCD method in this thesis is applied to explain the many subgroups of food at food inflation rate in Central Java into a more modest dimensions. The total proportion of the data variance can be explained by factors that are formed through a robust method of factor analysis in foodstuffs inflation data in Central Java that is equal to 72.9 percent larger than the classical factor analysis method which generates at 53.5 percent. This suggests that a more robust factor analysis method is able to cope with food inflation data in Central Java group containing outliers of the classical factor analysis method.