Sensitivity analysis

2826:(HDMR) (the term is due to H. Rabitz) is essentially an emulator approach, which involves decomposing the function output into a linear combination of input terms and interactions of increasing dimensionality. The HDMR approach exploits the fact that the model can usually be well-approximated by neglecting higher-order interactions (second or third-order and above). The terms in the truncated series can then each be approximated by e.g. polynomials or splines (REFS) and the response expressed as the sum of the main effects and interactions up to the truncation order. From this perspective, HDMRs can be seen as emulators which neglect high-order interactions; the advantage is that they are able to emulate models with higher dimensionality than full-order emulators. 1009:. This appears a logical approach as any change observed in the output will unambiguously be due to the single variable changed. Furthermore, by changing one variable at a time, one can keep all other variables fixed to their central or baseline values. This increases the comparability of the results (all 'effects' are computed with reference to the same central point in space) and minimizes the chances of computer program crashes, more likely when several input factors are changed simultaneously. OAT is frequently preferred by modelers because of practical reasons. In case of model failure under OAT analysis the modeler immediately knows which is the input factor responsible for the failure. 2869:

motivations of its author may become a matter of great importance, and a pure sensitivity analysis – with its emphasis on parametric uncertainty – may be seen as insufficient. The emphasis on the framing may derive inter-alia from the relevance of the policy study to different constituencies that are characterized by different norms and values, and hence by a different story about 'what the problem is' and foremost about 'who is telling the story'. Most often the framing includes more or less implicit assumptions, which could be political (e.g. which group needs to be protected) all the way to technical (e.g. which variable can be treated as a constant).

792: 2873:

quantitative information with the generation of `Pedigrees' of numbers. Sensitivity auditing has been especially designed for an adversarial context, where not only the nature of the evidence, but also the degree of certainty and uncertainty associated to the evidence, will be the subject of partisan interests. Sensitivity auditing is recommended in the European Commission guidelines for impact assessment, as well as in the report Science Advice for Policy by European Academies.

183: 2905:" I have proposed a form of organized sensitivity analysis that I call 'global sensitivity analysis' in which a neighborhood of alternative assumptions is selected and the corresponding interval of inferences is identified. Conclusions are judged to be sturdy only if the neighborhood of assumptions is wide enough to be credible and the corresponding interval of inferences is narrow enough to be useful." 2854:. In a design of experiments, one studies the effect of some process or intervention (the 'treatment') on some objects (the 'experimental units'). In sensitivity analysis one looks at the effect of varying the inputs of a mathematical model on the output of the model itself. In both disciplines one strives to obtain information from the system with a minimum of physical or numerical experiments. 2433:

thus overcomes the scale issue of traditional sensitivity analysis methods. More importantly, VARS is able to provide relatively stable and statistically robust estimates of parameter sensitivity with much lower computational cost than other strategies (about two orders of magnitude more efficient). Noteworthy, it has been shown that there is a theoretical link between the VARS framework and the

22: 1064:. With 5 inputs, the explored space already drops to less than 1% of the total parameter space. And even this is an overestimate, since the off-axis volume is not actually being sampled at all. Compare this to random sampling of the space, where the convex hull approaches the entire volume as more points are added. While the sparsity of OAT is theoretically not a concern for 2509:, that approximates the input/output behavior of the model itself. In other words, it is the concept of "modeling a model" (hence the name "metamodel"). The idea is that, although computer models may be a very complex series of equations that can take a long time to solve, they can always be regarded as a function of their inputs 2736:, although random designs can also be used, at the loss of some efficiency. The selection of the metamodel type and the training are intrinsically linked since the training method will be dependent on the class of metamodel. Some types of metamodels that have been used successfully for sensitivity analysis include: 571:-function) multiple times. Depending on the complexity of the model there are many challenges that may be encountered during model evaluation. Therefore, the choice of method of sensitivity analysis is typically dictated by a number of problem constraints, settings or challenges. Some of the most common are: 2910:

Note Leamer's emphasis is on the need for 'credibility' in the selection of assumptions. The easiest way to invalidate a model is to demonstrate that it is fragile with respect to the uncertainty in the assumptions or to show that its assumptions have not been taken 'wide enough'. The same concept is

1226:

function. Similar to OAT, local methods do not attempt to fully explore the input space, since they examine small perturbations, typically one variable at a time. It is possible to select similar samples from derivative-based sensitivity through Neural Networks and perform uncertainty quantification.

1225:

indicates that the derivative is taken at some fixed point in the space of the input (hence the 'local' in the name of the class). Adjoint modelling and Automated Differentiation are methods which allow to compute all partial derivatives at a cost at most 4-6 times of that for evaluating the original

782:

The various types of "core methods" (discussed below) are distinguished by the various sensitivity measures which are calculated. These categories can somehow overlap. Alternative ways of obtaining these measures, under the constraints of the problem, can be given. In addition, an engineering view of

2932:

Different statistical tests and measures are applied to the problem and different factors rankings are obtained. The test should instead be tailored to the purpose of the analysis, e.g. one uses Monte Carlo filtering if one is interested in which factors are most responsible for generating high/low

2872:

In order to take these concerns into due consideration the instruments of SA have been extended to provide an assessment of the entire knowledge and model generating process. This approach has been called 'sensitivity auditing'. It takes inspiration from NUSAP, a method used to qualify the worth of

2432:

for a given perturbation scale can be considered as a comprehensive illustration of sensitivity information, through linking variogram analysis to both direction and perturbation scale concepts. As a result, the VARS framework accounts for the fact that sensitivity is a scale-dependent concept, and

704:

Sometimes it is not possible to evaluate the code at all desired points, either because the code is confidential or because the experiment is not reproducible. The code output is only available for a given set of points, and it can be difficult to perform a sensitivity analysis on a limited set of

186:

Figure 1. Schematic representation of uncertainty analysis and sensitivity analysis. In mathematical modeling, uncertainty arises from a variety of sources - errors in input data, parameter estimation and approximation procedure, underlying hypothesis, choice of model, alternative model structures

2868:

It may happen that a sensitivity analysis of a model-based study is meant to underpin an inference, and to certify its robustness, in a context where the inference feeds into a policy or decision-making process. In these cases the framing of the analysis itself, its institutional context, and the

1249:

as direct measures of sensitivity. The regression is required to be linear with respect to the data (i.e. a hyperplane, hence with no quadratic terms, etc., as regressors) because otherwise it is difficult to interpret the standardised coefficients. This method is therefore most suitable when the

1020:

The proportion of input space which remains unexplored with an OAT approach grows superexponentially with the number of inputs. For example, a 3-variable parameter space which is explored one-at-a-time is equivalent to taking points along the x, y, and z axes of a cube centered at the origin. The

695:

A code is said to be stochastic when, for several evaluations of the code with the same inputs, different outputs are obtained (as opposed to a deterministic code when, for several evaluations of the code with the same inputs, the same output is always obtained). In this case, it is necessary to

2635:

to within an acceptable margin of error. Then, sensitivity measures can be calculated from the metamodel (either with Monte Carlo or analytically), which will have a negligible additional computational cost. Importantly, the number of model runs required to fit the metamodel can be orders of

550:

Taking into account uncertainty arising from different sources, whether in the context of uncertainty analysis or sensitivity analysis (for calculating sensitivity indices), requires multiple samples of the uncertain parameters and, consequently, running the model (evaluating the

121:, errors in input data, parameter estimation and approximation procedure, absence of information and poor or partial understanding of the driving forces and mechanisms, choice of underlying hypothesis of model, and so on. This uncertainty limits our confidence in the 745:

There are a large number of approaches to performing a sensitivity analysis, many of which have been developed to address one or more of the constraints discussed above. They are also distinguished by the type of sensitivity measure, be it based on (for example)

688:

is a vector or function. When outputs are correlated, it does not preclude the possibility of performing different sensitivity analyses for each output of interest. However, for models in which the outputs are correlated, the sensitivity measures can be hard to

84:

or system (numerical or otherwise) can be divided and allocated to different sources of uncertainty in its inputs. This involves estimating sensitivity indices that quantify the influence of an input or group of inputs on the output. A related practice is

1229:

One advantage of the local methods is that it is possible to make a matrix to represent all the sensitivities in a system, thus providing an overview that cannot be achieved with global methods if there is a large number of input and output variables.

1276:, and decompose the output variance into parts attributable to input variables and combinations of variables. The sensitivity of the output to an input variable is therefore measured by the amount of variance in the output caused by that input. 1884:

Variance-based methods allow full exploration of the input space, accounting for interactions, and nonlinear responses. For these reasons they are widely used when it is feasible to calculate them. Typically this calculation involves the use of

2954:

The importance of understanding and managing uncertainty in model results has inspired many scientists from different research centers all over the world to take a close interest in this subject. National and international agencies involved in

2468:

and represent the average marginal contribution of a given factors across all possible combinations of factors. These value are related to Sobol’s indices as their value falls between the first order Sobol’ effect and the total order effect.

109:(for example in biology, climate change, economics, renewable energy, agronomy...) can be highly complex, and as a result, its relationships between inputs and outputs may be faultily understood. In such cases, the model can be viewed as a 1012:

Despite its simplicity however, this approach does not fully explore the input space, since it does not take into account the simultaneous variation of input variables. This means that the OAT approach cannot detect the presence of

132:

In models involving many input variables, sensitivity analysis is an essential ingredient of model building and quality assurance and can be useful to determine the impact of a uncertain variable for a range of purposes, including:

2455:

to represent a multivariate function (the model) in the frequency domain, using a single frequency variable. Therefore, the integrals required to calculate sensitivity indices become univariate, resulting in computational savings.

1216: 1082:

Named after statistician Max D. Morris this method is suitable for screening systems with many parameters. This is also known as method of elementary effects because it combines repeated steps along the various parametric axes.

2834:

Sensitivity analysis via Monte Carlo filtering is also a sampling-based approach, whose objective is to identify regions in the space of the input factors corresponding to particular values (e.g., high or low) of the output.

2427:

Basically, the higher the variability the more heterogeneous is the response surface along a particular direction/parameter, at a specific perturbation scale. Accordingly, in the VARS framework, the values of directional

2489:). Generally, these methods focus on efficiently (by creating a metamodel of the costly function to be evaluated and/or by “ wisely ” sampling the factor space) calculating variance-based measures of sensitivity. 5002:

Van der Sluijs, JP; Craye, M; Funtowicz, S; Kloprogge, P; Ravetz, J; Risbey, J (2005). "Combining quantitative and qualitative measures of uncertainty in model based environmental assessment: the NUSAP system".

4001:

Kabir HD, Khosravi A, Nahavandi D, Nahavandi S. Uncertainty Quantification Neural Network from Similarity and Sensitivity. In2020 International Joint Conference on Neural Networks (IJCNN) 2020 Jul 19 (pp. 1-8).

4844:

Li, G.; Hu, J.; Wang, S.-W.; Georgopoulos, P.; Schoendorf, J.; Rabitz, H. (2006). "Random Sampling-High Dimensional Model Representation (RS-HDMR) and orthogonality of its different order component functions".

1852: 921:

using scatter plots, and observe the behavior of these pairs. The diagrams give an initial idea of the correlation and which input has an impact on the output. Figure 2 shows an example where two inputs,

1728: 585:

Time-consuming models are very often encountered when complex models are involved. A single run of the model takes a significant amount of time (minutes, hours or longer). The use of statistical model (

4524:

Storlie, C.B.; Swiler, L.P.; Helton, J.C.; Sallaberry, C.J. (2009). "Implementation and evaluation of nonparametric regression procedures for sensitivity analysis of computationally demanding models".

3651: 2945:

This is when one performs sensitivity analysis on one sub-model at a time. This approach is non conservative as it might overlook interactions among factors in different sub-models (Type II error).

2477:

The principle is to project the function of interest onto a basis of orthogonal polynomials. The Sobol indices are then expressed analytically in terms of the coefficients of this decomposition.

1458: 2422: 762:

Quantify the uncertainty in each input (e.g. ranges, probability distributions). Note that this can be difficult and many methods exist to elicit uncertainty distributions from subjective data.

147:

Uncertainty reduction, through the identification of model input that cause significant uncertainty in the output and should therefore be the focus of attention in order to increase robustness.

3006: 779:

In some cases this procedure will be repeated, for example in high-dimensional problems where the user has to screen out unimportant variables before performing a full sensitivity analysis.

2485:

A number of methods have been developed to overcome some of the constraints discussed above, which would otherwise make the estimation of sensitivity measures infeasible (most often due to

2926:, although even then it may be hard to build distributions with great confidence. The subjectivity of the probability distributions or ranges will strongly affect the sensitivity analysis. 2285:(CDFs) to characterize the maximum distance between the unconditional output distribution and conditional output distribution (obtained by varying all input parameters and by setting the 2716:"Training" the metamodel using the sample data from the model – this generally involves adjusting the metamodel parameters until the metamodel mimics the true model as well as possible. 5101: 2633: 2798:, in conjunction with canonical models such as noisy models. Noisy models exploit information on the conditional independence between variables to significantly reduce dimensionality. 2321:

One of the major shortcomings of the previous sensitivity analysis methods is that none of them considers the spatially ordered structure of the response surface/output of the model

4397:

Haghnegahdar, Amin; Razavi, Saman (September 2017). "Insights into sensitivity analysis of Earth and environmental systems models: On the impact of parameter perturbation scale".

433: 317: 737:

To address the various constraints and challenges, a number of methods for sensitivity analysis have been proposed in the literature, which we will examine in the next section.

2991: 2117: 2675: 2580: 2360:

and covariograms, variogram analysis of response surfaces (VARS) addresses this weakness through recognizing a spatially continuous correlation structure to the values of

1518: 2278:-importance measure, the new correlation coefficient of Chatterjee, the Wasserstein correlation of Wiesel and the kernel-based sensitivity measures of Barr and Rabitz. 2190: 2082: 2249: 1487: 990:

One of the simplest and most common approaches is that of changing one-factor-at-a-time (OAT), to see what effect this produces on the output. OAT customarily involves

5368: 1996: 2276: 380: 3011: 2542: 2354: 2148: 1969: 1879: 1628: 1575: 1548: 1370: 1308: 1140: 974: 947: 899: 520: 2704: 1062: 1889:

methods, but since this can involve many thousands of model runs, other methods (such as metamodels) can be used to reduce computational expense when necessary.

2378: 2303: 2214: 1939: 1915: 1601: 1328: 1113: 919: 731: 686: 615: 569: 540: 477: 453: 340: 249: 217: 3114:

Saltelli, A.; Ratto, M.; Andreas, T.; Campolongo, F.; Gariboni, J.; Gatelli, D.; Saisana, M.; Tarantola, S. (2008). "Global sensitivity analysis: the primer".

187:

and so on. They propagate through the model and have an impact on the output. The uncertainty on the output is described via uncertainty analysis (represented

125:

of the model's response or output. Further, models may have to cope with the natural intrinsic variability of the system (aleatory), such as the occurrence of

624:, which grows exponentially in size with the number of inputs. Therefore, screening methods can be useful for dimension reduction. Another way to tackle the 2255:, the moment-independent global sensitivity measure satisfies zero-independence. This is a relevant statistical property also known as Renyi's postulate D. 1148: 3001: 1897:

Moment-independent methods extend variance-based techniques by considering the probability density or cumulative distribution function of the model output

4044:

Borgonovo, E., Tarantola, S., Plischke, E., Morris, M. D. (2014). "Transformations and invariance in the sensitivity analysis of computer experiments".

4666:"Categorical Inputs, Sensitivity Analysis, Optimization and Importance Tempering with tgp Version 2, an R Package for Treed Gaussian Process Models" 4433: 2996: 640:

between model inputs, but sometimes inputs can be strongly correlated. Correlations between inputs must then be taken into account in the analysis.

156:

Enhancing communication from modelers to decision makers (e.g. by making recommendations more credible, understandable, compelling or persuasive).

5112:

Science Advice for Policy by European Academies, Making sense of science for policy under conditions of complexity and uncertainty, Berlin, 2019.

1633: 173:

To identify important connections between observations, model inputs, and predictions or forecasts, leading to the development of better models.

1283:: they represent the proportion of variance explained by an input or group of inputs. This expression essentially measures the contribution of 153:

Model simplification – fixing model input that has no effect on the output, or identifying and removing redundant parts of the model structure.

765:

Identify the model output to be analysed (the target of interest should ideally have a direct relation to the problem tackled by the model).

3147:

Saltelli, A.; Tarantola, S.; Campolongo, F.; Ratto, M. (2004). "Sensitivity analysis in practice: a guide to assessing scientific models".

2847:

in the conclusions of the study, sensitivity analysis tries to identify what source of uncertainty weighs more on the study's conclusions.

1377: 159:

Finding regions in the space of input factors for which the model output is either maximum or minimum or meets some optimum criterion (see

2305:-th input, consequentially). The difference between the unconditional and conditional output distribution is usually calculated using the 799:(vertical axis) is a function of four factors. The points in the four scatterplots are always the same though sorted differently, i.e. by 4801:

Cardenas, IC (2019). "On the use of Bayesian networks as a meta-modeling approach to analyse uncertainties in slope stability analysis".

3553: 2939:

This may be acceptable for the quality assurance of sub-models but should be avoided when presenting the results of the overall analysis.

5224: 1733: 4551:

Wang, Shangying; Fan, Kai; Luo, Nan; Cao, Yangxiaolu; Wu, Feilun; Zhang, Carolyn; Heller, Katherine A.; You, Lingchong (2019-09-25).

3742:

Bailis, R.; Ezzati, M.; Kammen, D. (2005). "Mortality and Greenhouse Gas Impacts of Biomass and Petroleum Energy Futures in Africa".

2972: 5193: 2281:

Another measure for global sensitivity analysis, in the category of moment-independent approaches, is the PAWN index. It relies on

1029:

which has a volume only 1/6th of the total parameter space. More generally, the convex hull of the axes of a hyperrectangle forms a

2888:

In uncertainty and sensitivity analysis there is a crucial trade off between how scrupulous an analyst is in exploring the input

1998:, can be defined through an equation similar to variance-based indices replacing the conditional expectation with a distance, as 3224:

Bahremand, A.; De Smedt, F. (2008). "Distributed Hydrological Modeling and Sensitivity Analysis in Torysa Watershed, Slovakia".

113:, i.e. the output is an "opaque" function of its inputs. Quite often, some or all of the model inputs are subject to sources of 3795:

Murphy, J.; et al. (2004). "Quantification of modelling uncertainties in a large ensemble of climate change simulations".

4766:

Ratto, M.; Pagano, A. (2010). "Using recursive algorithms for the efficient identification of smoothing spline ANOVA models".

5064:

Lo Piano, S; Robinson, M (2019). "Nutrition and public health economic evaluations under the lenses of post normal science".

3669: 3131: 3039: 2823: 2788: 2446: 2309:(KS). The PAWN index for a given input parameter is then obtained by calculating the summary statistics over all KS values. 5358: 2383: 3084: 2434: 1263: 874:

The first intuitive approach (especially useful in less complex cases) is to analyze the relationship between each input

747: 655: 542:(by calculating the corresponding sensitivity indices). Figure 1 provides a schematic representation of this statement. 2976: 2748: 4445: 3601: 3034: 2544:. By running the model at a number of points in the input space, it may be possible to fit a much simpler metamodels 2282: 637: 57: 1268:

Variance-based methods are a class of probabilistic approaches which quantify the input and output uncertainties as

3461: 3410: 492: 2968: 578:

Sensitivity analysis is almost always performed by running the model a (possibly large) number of times, i.e. a

5363: 2585: 783:

the methods that takes into account the four important sensitivity analysis parameters has also been proposed.

2636:

magnitude less than the number of runs required to directly estimate the sensitivity measures from the model.

5348: 4619:

Oakley, J.; O'Hagan, A. (2004). "Probabilistic sensitivity analysis of complex models: a Bayesian approach".

3892: 2811: 620:

The model has a large number of uncertain inputs. Sensitivity analysis is essentially the exploration of the

347: 3364:"Quasi-Monte Carlo technique in global sensitivity analysis of wind resource assessment with a study on UAE" 696:

separate the variability of the output due to the variability of the inputs from that due to stochasticity.

3318:

Effective Groundwater Model Calibration, with Analysis of Data, Sensitivities, Predictions, and Uncertainty

1251: 488: 195:

showing the proportion that each source of uncertainty contributes to the total uncertainty of the output).

188: 4553:"Massive computational acceleration by using neural networks to emulate mechanism-based biological models" 4268:"A simple and efficient method for global sensitivity analysis based on cumulative distribution functions" 1630:

and its interactions with any of the other input variables. The total effect index is given as following:

3406:"Generalized Hoeffding-Sobol decomposition for dependent variables - application to sensitivity analysis" 3064: 2306: 997:

returning the variable to its nominal value, then repeating for each of the other inputs in the same way.

985: 387: 258: 2843:

Sensitivity analysis is closely related with uncertainty analysis; while the latter studies the overall

1254:

is large. The advantages of regression analysis are that it is simple and has a low computational cost.

150:

Searching for errors in the model (by encountering unexpected relationships between inputs and outputs).

3079: 2771:, where a succession of simple regressions are used to weight data points to sequentially reduce error. 114: 94: 90: 4974:

Hornberger, G.; Spear, R. (1981). "An approach to the preliminary analysis of environmental systems".

4350:"A new framework for comprehensive, robust, and efficient global sensitivity analysis: 2. Application" 2087: 4489: 4239:

Barr, J., Rabitz, H. (31 March 2022). "A Generalized Kernel Method for Global Sensitivity Analysis".

3548: 3029: 2959:

studies have included sections devoted to sensitivity analysis in their guidelines. Examples are the

2721: 2710:

Sampling (running) the model at a number of points in its input space. This requires a sample design.

755: 144:

Increased understanding of the relationships between input and output variables in a system or model.

32: 191:

on the output) and their relative importance is quantified via sensitivity analysis (represented by

170:

For calibration of models with large number of parameters, by focusing on the sensitive parameters.

5153: 5124: 3512: 3195:"Sensitivity Analysis of Normative Economic Models: Theoretical Framework and Practical Strategies" 3044: 2733: 2642: 2547: 1273: 1246: 1014: 122: 5250:

Pianosi, F.; Beven, K.; Freer, J.; Hall, J.W.; Rougier, J.; Stephenson, D.B.; Wagener, T. (2016).

4633: 1492: 4665: 4301:"A new framework for comprehensive, robust, and efficient global sensitivity analysis: 1. Theory" 2193: 2153: 1886: 625: 118: 4992:

Box GEP, Hunter WG, Hunter, J. Stuart. Statistics for experimenters . New York: Wiley & Sons

2001: 4628: 2778: 2222: 1463: 3507: 1974: 39: 3054: 2851: 2486: 2261: 791: 769: 579: 5245: 3921:

Morris, M. D. (1991). "Factorial Sampling Plans for Preliminary Computational Experiments".

2512: 2324: 5263: 5252:"Sensitivity analysis of environmental models: A systematic review with practical workflow" 5012: 4932: 4897: 4854: 4810: 4713: 4564: 4406: 4361: 4312: 3806: 3753: 3700: 3332: 3267:

Hill, M.; Kavetski, D.; Clark, M.; Ye, M.; Arabi, M.; Lu, D.; Foglia, L.; Mehl, S. (2015).

3233: 3059: 2863: 2850:

The problem setting in sensitivity analysis also has strong similarities with the field of

2252: 2126: 2120: 1947: 1918: 1857: 1606: 1553: 1526: 1348: 1286: 1118: 952: 925: 877: 498: 456: 86: 2680: 1036: 8: 5353: 5200: 3089: 2960: 1238: 1002: 751: 5267: 5016: 4936: 4901: 4858: 4814: 4717: 4568: 4410: 4365: 4316: 4000: 3810: 3757: 3704: 3237: 2505:

approaches that involve building a relatively simple mathematical function, known as an

5244:

International Series in Management Science and Operations Research, Springer New York.

5162: 5133: 5081: 5046: 4956: 4826: 4783: 4646: 4593: 4552: 4203: 4158: 3938: 3868: 3830: 3777: 3724: 3562: 3470: 3419: 3386: 3249: 3210: 2987:

The following pages discuss sensitivity analyses in relation to specific applications:

2923: 2752: 2740: 2363: 2288: 2199: 1924: 1900: 1586: 1313: 1098: 1092: 904: 716: 671: 600: 554: 525: 462: 438: 325: 234: 202: 164: 106: 81: 4194:

Wiesel, J. C. W. (November 2022). "Measuring association with Wasserstein distances".

3549:"An efficient methodology for modeling complex computer codes with Gaussian processes" 2810:. In all cases, it is useful to check the accuracy of the emulator, for example using 1211:{\displaystyle \left|{\frac {\partial Y}{\partial X_{i}}}\right|_{{\textbf {x}}^{0}},} 5218: 5085: 5038: 5024: 4948: 4870: 4830: 4642: 4598: 4580: 4441: 4379: 4330: 4221: 4176: 4131: 4096: 4061: 3822: 3769: 3744: 3728: 3716: 3691: 3665: 3597: 3529: 3488: 3437: 3390: 3298: 3290: 3194: 3127: 3094: 2964: 2956: 2795: 2768: 2258:

The class of moment-independent sensitivity measures includes indicators such as the

1242: 1006: 710: 647: 590: 5050: 4960: 4739:

Sudret, B. (2008). "Global sensitivity analysis using polynomial chaos expansions".

3781: 3166:

Der Kiureghian, A.; Ditlevsen, O. (2009). "Aleatory or epistemic? Does it matter?".

1250:

model response is in fact linear; linearity can be confirmed, for instance, if the

5333: 5281: 5271: 5073: 5028: 5020: 4940: 4905: 4862: 4818: 4787: 4775: 4748: 4721: 4680: 4650: 4638: 4588: 4572: 4533: 4501: 4470: 4414: 4369: 4320: 4279: 4248: 4213: 4168: 4123: 4088: 4053: 4043: 3930: 3901: 3860: 3848: 3814: 3797: 3761: 3708: 3657: 3630: 3572: 3521: 3480: 3429: 3378: 3363: 3344: 3280: 3253: 3241: 3206: 3179: 3175: 3119: 2897: 2803: 2784: 2774: 2502: 2480: 994:

moving one input variable, keeping others at their baseline (nominal) values, then,

651: 5276: 5251: 5077: 4822: 4418: 4284: 4267: 4172: 3834: 3687:"Photosynthetic Control of Atmospheric Carbonyl Sulfide During the Growing Season" 3456: 3405: 775:

Using the resulting model outputs, calculate the sensitivity measures of interest.

713:) from the available data (that we use for training) to approximate the code (the 495:,...), sensitivity analysis aims to measure and quantify the impact of each input 5316:. Mathematics in Science and Engineering, 177. Academic Press, Inc., Orlando, FL. 4944: 3591: 3508:"Global sensitivity analysis of stochastic computer models with joint metamodels" 3069: 1269: 1030: 138: 4888:

Li, G. (2002). "Practical approaches to construct RS-HDMR component functions".

4803:

Georisk: Assessment and Management of Risk for Engineered Systems and Geohazards

4725: 3988:

Evaluating Derivatives, Principles and Techniques of Algorithmic Differentiation

3888:"Threshold for the volume spanned by random points with independent coordinates" 2889: 2497:

Metamodels (also known as emulators, surrogate models or response surfaces) are

5321:

Ecosystem Modeling in Theory and Practice: An Introduction with Case Histories.

4752: 4576: 4537: 4505: 4202:(4). Bernoulli Society for Mathematical Statistics and Probability: 2816–2832. 4127: 3686: 3576: 3348: 2762: 2725: 2452: 484: 4779: 3906: 3887: 3525: 3245: 1001:

Sensitivity may then be measured by monitoring changes in the output, e.g. by

827:

in turn. Note that the abscissa is different for each plot: (−5, +5) for

5342: 4584: 4383: 4334: 4225: 4180: 4135: 4100: 4065: 4025:

Sobol', I (1993). "Sensitivity analysis for non-linear mathematical models".

4013:

Sobol', I (1990). "Sensitivity estimates for nonlinear mathematical models".

3533: 3492: 3441: 2913:

uncertainties in inputs must be suppressed lest outputs become indeterminate.

2758: 2729: 2498: 2465: 1077: 4046:

Journal of the Royal Statistical Society. Series B (Statistical Methodology)

3765: 3712: 3661: 3635: 3618: 2747:), where any combination of output points is assumed to be distributed as a 5309: 5042: 4952: 4874: 4602: 3826: 3773: 3720: 3333:"Survey of sampling based methods for uncertainty and sensitivity analysis" 3302: 3049: 1065: 160: 5296:

Useless Arithmetic. Why Environmental Scientists Can't Predict the Future.

4685: 3123: 4374: 4349: 4325: 4300: 3973:

Sensitivity and Uncertainty Analysis: Applications to Large-Scale Systems

3649: 2920:

Not enough information to build probability distributions for the inputs:

2844: 1022: 97:; ideally, uncertainty and sensitivity analysis should be run in tandem. 77: 3818: 3331:

Helton, J. C.; Johnson, J. D.; Salaberry, C. J.; Storlie, C. B. (2006).

758:. In general, however, most procedures adhere to the following outline: 5286: 5166: 5137: 5033: 4252: 4217: 4092: 4057: 3942: 3872: 3285: 3268: 2806:

problem, which can be difficult if the response of the model is highly

1026: 665: 480: 126: 4909: 4866: 4474: 3484: 3433: 3382: 3294: 5001: 4149:

Chatterjee, S. (2 October 2021). "A New Coefficient of Correlation".

3074: 3024: 2893: 2807: 2429: 2357: 621: 228: 192: 182: 110: 5122:

Leamer, Edward E. (1983). "Let's Take the Con Out of Econometrics".

5102:

European Commission. 2021. “Better Regulation Toolbox.” November 25.

3934: 3864: 2677:(metamodel) that is a sufficiently close approximation to the model 70:

Study of uncertainty in the output of a mathematical model or system

5242:

Sensitivity Analysis: An Introduction for the Management Scientist.

4701: 4208: 4163: 4114:

Borgonovo, E. (June 2007). "A new uncertainty importance measure".

2751:. Recently, "treed" Gaussian processes have been used to deal with 1330:(averaged over variations in other variables), and is known as the 141:

of the results of a model or system in the presence of uncertainty.

4461:

Owen, A. B. (1 January 2014). "Sobol' Indices and Shapley Value".

3567: 3475: 3457:"Sensitivity analysis for multidimensional and functional outputs" 3424: 650:, can inaccurately measure sensitivity when the model response is 594: 2744: 2356:

in the parameter space. By utilizing the concepts of directional

2312: 3269:"Practical use of computationally frugal model analysis methods" 3146: 2982: 2817: 2481:

Complementary research approaches for time-consuming simulations

1847:{\displaystyle X_{\sim i}=(X_{1},...,X_{i-1},X_{i+1},...,X_{p})} 1017:

between input variables and is unsuitable for nonlinear models.

4702:"Bayesian sensitivity analysis of bifurcating nonlinear models" 4490:"Global sensitivity analysis using polynomial chaos expansions" 3971:

Cacuci, Dan G.; Ionescu-Bujor, Mihaela; Navon, Michael (2005).

1520:

denote the variance and expected value operators respectively.

795:

Figure 2. Sampling-based sensitivity analysis by scatterplots.

5334:

Web site with material from SAMO conference series (1995-2025)

4469:(1). Society for Industrial and Applied Mathematics: 245–251. 3617:

Sacks, J.; Welch, W. J.; Mitchell, T. J.; Wynn, H. P. (1989).

3499: 3113: 5151:

Leamer, Edward E. (1985). "Sensitivity Analyses Would Help".

4523: 3448: 3330: 2911:

expressed by Jerome R. Ravetz, for whom bad modeling is when

1241:, in the context of sensitivity analysis, involves fitting a 772:, dictated by the method of choice and the input uncertainty. 646:

Some sensitivity analysis approaches, such as those based on

5323:

John Wiley & Sons, New York, NY. isbn=978-0-471-34165-9.

4247:(1). Society for Industrial and Applied Mathematics: 27–54. 3851:(1999). "One-Factor-at-a-Time Versus Designed Experiments". 2896:

may be. The point is well illustrated by the econometrician

2713:

Selecting a type of emulator (mathematical function) to use.

2440: 4079:

Rényi, A. (1 September 1959). "On measures of dependence".

3546: 1723:{\displaystyle S_{i}^{T}=1-{\frac {V(\mathbb {E} )}{V(Y)}}} 199:

The object of study for sensitivity analysis is a function

3547:

Marrel, A.; Iooss, B.; Van Dorpe, F.; Volkova, E. (2008).

3505: 2881:

Some common difficulties in sensitivity analysis include:

1577:

has with other variables. A further measure, known as the

545: 177: 3506:

Marrel, A.; Iooss, B.; Da Veiga, S.; Ribatet, M. (2012).

2639:

Clearly, the crux of an metamodel approach is to find an

668:, sensitivity analysis extends to cases where the output 582:-based approach. This can be a significant problem when: 4923:

Rabitz, H (1989). "System analysis at molecular scale".

3970: 3650:

Da Veiga, S., Gamboa, F., Iooss, B., Prieur, C. (2021).

1550:

does not measure the uncertainty caused by interactions

4436:. In Petropoulos, George; Srivastava, Prashant (eds.). 4434:"Challenges and Future Outlook of Sensitivity Analysis" 3454: 3165: 2451:

The Fourier amplitude sensitivity test (FAST) uses the

628:

is to use sampling based on low discrepancy sequences.

617:-function is one way of reducing the computation costs. 4265: 3616: 3593:

Uncertain Judgements: Eliciting Experts' Probabilities

3455:

Gamboa, F.; Janon, A.; Klein, T.; Lagnoux, A. (2014).

522:

or a group of inputs on the variability of the output

5249: 3885: 3403: 3309: 2683: 2645: 2588: 2550: 2515: 2386: 2366: 2327: 2291: 2264: 2225: 2202: 2156: 2129: 2090: 2004: 1977: 1950: 1927: 1903: 1860: 1736: 1636: 1609: 1589: 1556: 1529: 1495: 1466: 1453:{\displaystyle S_{i}={\frac {V(\mathbb {E} )}{V(Y)}}} 1380: 1351: 1316: 1289: 1151: 1121: 1101: 1039: 955: 928: 907: 880: 719: 674: 603: 557: 528: 501: 465: 441: 390: 350: 328: 261: 237: 205: 4843: 4027:

Mathematical Modeling & Computational Experiment

2417:{\displaystyle {\frac {\partial Y}{\partial x_{i}}}} 5301:Santner, T. J.; Williams, B. J.; Notz, W.I. (2003) 4438:

Sensitivity Analysis in Earth Observation Modelling

3266: 4699: 4396: 4238: 3741: 2922:Probability distributions can be constructed from 2698: 2669: 2627: 2574: 2536: 2416: 2372: 2348: 2297: 2270: 2243: 2208: 2184: 2142: 2111: 2076: 1990: 1963: 1933: 1909: 1873: 1846: 1722: 1622: 1595: 1569: 1542: 1512: 1481: 1452: 1364: 1322: 1302: 1210: 1134: 1107: 1091:Local derivative-based methods involve taking the 1056: 968: 941: 913: 893: 725: 680: 609: 563: 534: 514: 471: 447: 427: 374: 334: 311: 243: 211: 5369:Mathematical and quantitative methods (economics) 5314:Design sensitivity analysis of structural systems 4618: 3223: 1086: 5340: 5063: 4348:Razavi, Saman; Gupta, Hoshin V. (January 2016). 4299:Razavi, Saman; Gupta, Hoshin V. (January 2016). 4193: 4081:Acta Mathematica Academiae Scientiarum Hungarica 636:Most common sensitivity analysis methods assume 459:aims to describe the distribution of the output 4973: 4151:Journal of the American Statistical Association 3361: 1944:The moment-independent sensitivity measures of 1279:This amount is quantified and calculated using 740: 4463:SIAM/ASA Journal on Uncertainty Quantification 4241:SIAM/ASA Journal on Uncertainty Quantification 4148: 3540: 3404:Chastaing, G.; Gamboa, F.; Prieur, C. (2012). 2949: 1854:denotes the set of all input variables except 1523:Importantly, first-order sensitivity index of 4550: 4113: 4107: 4037: 3920: 3619:"Design and Analysis of Computer Experiments" 3260: 2983:Specific applications of sensitivity analysis 2818:High-dimensional model representations (HDMR) 1892: 4663: 4440:(1st ed.). Elsevier. pp. 397–415. 3958:Sensitivity and Uncertainty Analysis: Theory 3554:Computational Statistics & Data Analysis 3315: 2876: 1917:. Thus, they do not refer to any particular 1681: 1414: 5303:Design and Analysis of Computer Experiments 5294:Pilkey, O. H. and L. Pilkey-Jarvis (2007), 5097: 5095: 4765: 4741:Reliability Engineering & System Safety 4700:Becker, W.; Worden, K.; Rowson, J. (2013). 4526:Reliability Engineering & System Safety 4460: 4431: 4347: 4298: 4116:Reliability Engineering & System Safety 3589: 3371:Journal of Renewable and Sustainable Energy 768:Run the model a number of times using some 654:with respect to its inputs. In such cases, 4481: 4142: 4024: 4012: 3107: 2720:Sampling the model can often be done with 5285: 5275: 5057: 5032: 4986: 4684: 4632: 4592: 4519: 4517: 4515: 4373: 4324: 4283: 4207: 4162: 4078: 3905: 3653:Basics and Trends in Sensitivity Analysis 3634: 3566: 3474: 3423: 3337:Reliability Engineering and System Safety 3284: 2973:Intergovernmental Panel on Climate Change 2628:{\displaystyle {\hat {f}}(X)\approx f(X)} 2441:Fourier amplitude sensitivity test (FAST) 2313:Variogram analysis of response surfaces ( 1671: 1497: 1404: 1257: 705:data. We then build a statistical model ( 58:Learn how and when to remove this message 5183:, New Internationalist Publications Ltd. 5092: 4800: 4706:Mechanical Systems and Signal Processing 4157:(536). Taylor & Francis: 2009–2022. 3994: 3985: 3684: 3362:Tsvetkova, O.; Ouarda, T.B.M.J. (2019). 3217: 2829: 976:are highly correlated with the output. 790: 181: 4272:Environmental Modelling \& Software 3847: 3192: 3186: 2857: 1372:, Sobol index is defined as following: 1310:alone to the uncertainty (variance) in 546:Challenges, settings and related issues 178:Mathematical formulation and vocabulary 5341: 5319:Hall, C. A. S. and Day, J. W. (1977). 5256:Environmental Modelling & Software 5223:: CS1 maint: archived copy as title ( 5150: 5121: 4922: 4738: 4614: 4612: 4512: 4487: 4399:Environmental Modelling & Software 4187: 3886:Gatzouras, D; Giannopoulos, A (2009). 3794: 3397: 2937:Too many model outputs are considered: 1233: 979: 4995: 4768:AStA Advances in Statistical Analysis 4664:Gramacy, R. B.; Taddy, M. A. (2010). 4072: 3140: 3040:Fourier amplitude sensitivity testing 2824:high-dimensional model representation 2789:high-dimensional model representation 2765:are trained, and the result averaged. 2706:. This requires the following steps, 2447:Fourier amplitude sensitivity testing 5298:New York: Columbia University Press. 4259: 4232: 3929:(2). Taylor & Francis: 161–174. 2802:The use of an emulator introduces a 2787:, normally used in conjunction with 2781:to approximate the response surface. 2472: 1247:standardized regression coefficients 1068:, true linearity is rare in nature. 384:The variability in input parameters 15: 4976:Journal of Environmental Management 4609: 4494:Bayesian Networks in Dependability] 3975:. Vol. II. Chapman & Hall. 3085:Variance-based sensitivity analysis 2838: 1264:Variance-based sensitivity analysis 1192: 428:{\displaystyle X_{i},i=1,\ldots ,p} 312:{\displaystyle X=(X_{1},...,X_{p})} 13: 5234: 4887: 3960:. Vol. I. Chapman & Hall. 3955: 3685:Campbell, J.; et al. (2008). 3211:10.1111/j.1574-0862.1997.tb00449.x 2977:US Environmental Protection Agency 2749:multivariate Gaussian distribution 2459: 2398: 2390: 2380:, and hence also to the values of 1168: 1160: 786: 14: 5380: 5327: 5308:Haug, Edward J.; Choi, Kyung K.; 4266:Pianosi, F., Wagener, T. (2015). 3590:O'Hagan, A.; et al. (2006). 3035:Experimental uncertainty analysis 2437:and derivative-based approaches. 2283:Cumulative Distribution Functions 862:is most important in influencing 5025:10.1111/j.1539-6924.2005.00604.x 4643:10.1111/j.1467-9868.2004.05304.x 3462:Electronic Journal of Statistics 3411:Electronic Journal of Statistics 2930:Unclear purpose of the analysis: 2112:{\displaystyle d(\cdot ,\cdot )} 1245:to the model response and using 1115:with respect to an input factor 1025:bounding all these points is an 20: 5186: 5173: 5144: 5115: 5106: 4967: 4916: 4881: 4847:Journal of Physical Chemistry A 4837: 4794: 4759: 4732: 4693: 4673:Journal of Statistical Software 4657: 4544: 4454: 4425: 4390: 4341: 4292: 4006: 3979: 3964: 3949: 3914: 3879: 3841: 3788: 3735: 3678: 3643: 3610: 3583: 3316:Hill, M.; Tiedeman, C. (2007). 2969:Office of Management and Budget 2791:(HDMR) truncations (see below). 1033:which has a volume fraction of 706: 662:Multiple or functional outputs: 586: 89:, which has a greater focus on 3355: 3324: 3180:10.1016/j.strusafe.2008.06.020 3159: 3007:Multi-criteria decision making 2693: 2687: 2664: 2658: 2652: 2622: 2616: 2607: 2601: 2595: 2569: 2563: 2557: 2531: 2525: 2343: 2337: 2232: 2229: 2166: 2123:between probability measures, 2106: 2094: 2071: 2068: 2052: 2027: 2021: 1841: 1753: 1714: 1708: 1700: 1697: 1675: 1667: 1583:, gives the total variance in 1507: 1501: 1476: 1470: 1444: 1438: 1430: 1427: 1408: 1400: 1087:Derivative-based local methods 866:as it imparts more 'shape' on 366: 360: 306: 268: 29:This article needs editing to 1: 5277:10.1016/j.envsoft.2016.02.008 5078:10.1016/j.futures.2019.06.008 4890:Journal of Physical Chemistry 4823:10.1080/17499518.2018.1498524 4419:10.1016/j.envsoft.2017.03.031 4285:10.1016/j.envsoft.2015.01.004 4173:10.1080/01621459.2020.1758115 4015:Matematicheskoe Modelirovanie 3893:Israel Journal of Mathematics 3101: 2963:(see e.g. the guidelines for 2761:, in which a large number of 2670:{\displaystyle {\hat {f}}(X)} 2575:{\displaystyle {\hat {f}}(X)} 2492: 1333:first-order sensitivity index 435:have an impact on the output 100: 5181:No-Nonsense Guide to Science 4945:10.1126/science.246.4927.221 4432:Gupta, H; Razavi, S (2016). 2755:and discontinuous responses. 1513:{\displaystyle \mathbb {E} } 1252:coefficient of determination 741:Sensitivity analysis methods 622:multidimensional input space 7: 5359:Business intelligence terms 4726:10.1016/j.ymssp.2012.05.010 4023:; translated in English in 3065:Probability bounds analysis 3017: 2950:SA in international context 2892:and how wide the resulting 2886:Assumptions vs. inferences: 2775:Polynomial chaos expansions 2185:{\displaystyle P_{Y|X_{i}}} 986:One-factor-at-a-time method 10: 5385: 4753:10.1016/j.ress.2007.04.002 4577:10.1038/s41467-019-12342-y 4538:10.1016/j.ress.2009.05.007 4506:10.1016/j.ress.2007.04.002 4128:10.1016/J.RESS.2006.04.015 3577:10.1016/j.csda.2008.03.026 3349:10.1016/j.ress.2005.11.017 3226:Water Resources Management 3080:Uncertainty quantification 2861: 2444: 2077:{\displaystyle \xi _{i}=E} 1893:Moment-independent methods 1261: 1075: 983: 342:, presented as following: 95:propagation of uncertainty 91:uncertainty quantification 4780:10.1007/s10182-010-0148-8 3907:10.1007/s11856-009-0007-z 3526:10.1007/s11222-011-9274-8 3246:10.1007/s11269-007-9168-x 3030:Elementary effects method 2877:Pitfalls and difficulties 2722:low-discrepancy sequences 2244:{\displaystyle d()\geq 0} 1482:{\displaystyle V(\cdot )} 1274:probability distributions 1071: 664:Generally introduced for 5154:American Economic Review 5125:American Economic Review 4354:Water Resources Research 4305:Water Resources Research 4122:(6). Elsevier: 771–784. 3513:Statistics and Computing 3320:. John Wiley & Sons. 3045:Info-gap decision theory 2979:'s modeling guidelines. 2734:Latin hypercube sampling 2464:Shapley effects rely on 1991:{\displaystyle \xi _{i}} 1272:, represented via their 76:is the study of how the 31:comply with Knowledge's 3766:10.1126/science.1106881 3713:10.1126/science.1164015 3662:10.1137/1.9781611976694 3193:Pannell, D. J. (1997). 2728:– due to mathematician 2307:Kolmogorov–Smirnov test 2271:{\displaystyle \delta } 2194:conditional probability 748:variance decompositions 656:variance-based measures 626:curse of dimensionality 375:{\displaystyle Y=f(X).} 5240:Borgonovo, E. (2017). 3199:Agricultural Economics 3116:John Wiley \& Sons 2992:Environmental sciences 2943:Piecewise sensitivity: 2907: 2779:orthogonal polynomials 2700: 2671: 2629: 2576: 2538: 2537:{\displaystyle Y=f(X)} 2418: 2374: 2350: 2349:{\displaystyle Y=f(X)} 2299: 2272: 2245: 2210: 2186: 2144: 2113: 2078: 1992: 1965: 1935: 1911: 1875: 1848: 1724: 1624: 1597: 1571: 1544: 1514: 1483: 1454: 1366: 1324: 1304: 1258:Variance-based methods 1212: 1136: 1109: 1058: 970: 943: 915: 895: 871: 841:, (−10, +10) for 727: 682: 611: 576:Computational expense: 565: 536: 516: 473: 449: 429: 376: 336: 313: 245: 213: 196: 5364:Mathematical modeling 4686:10.18637/jss.v033.i06 4557:Nature Communications 3986:Griewank, A. (2000). 3853:American Statistician 3636:10.1214/ss/1177012413 3596:. Chichester: Wiley. 3124:10.1002/9780470725184 3055:Perturbation analysis 2933:values of the output. 2903: 2852:design of experiments 2830:Monte Carlo filtering 2701: 2672: 2630: 2577: 2539: 2487:computational expense 2419: 2375: 2351: 2300: 2273: 2246: 2211: 2192:are the marginal and 2187: 2145: 2143:{\displaystyle P_{Y}} 2114: 2079: 1993: 1966: 1964:{\displaystyle X_{i}} 1936: 1912: 1876: 1874:{\displaystyle X_{i}} 1849: 1725: 1625: 1623:{\displaystyle X_{i}} 1598: 1572: 1570:{\displaystyle X_{i}} 1545: 1543:{\displaystyle X_{i}} 1515: 1484: 1455: 1367: 1365:{\displaystyle X_{i}} 1325: 1305: 1303:{\displaystyle X_{i}} 1213: 1137: 1135:{\displaystyle X_{i}} 1110: 1059: 971: 969:{\displaystyle Z_{4}} 944: 942:{\displaystyle Z_{3}} 916: 896: 894:{\displaystyle Z_{i}} 794: 770:design of experiments 728: 702:Data-driven approach: 683: 658:are more appropriate. 612: 566: 537: 517: 515:{\displaystyle X_{i}} 474: 450: 430: 377: 337: 314: 246: 214: 185: 165:Monte Carlo filtering 119:errors of measurement 5349:Sensitivity analysis 5179:Ravetz, J.R., 2007, 4375:10.1002/2015WR017559 4326:10.1002/2015WR017558 3343:(10–11): 1175–1209. 3149:Wiley Online Library 3060:Probabilistic design 2864:Sensitivity auditing 2858:Sensitivity auditing 2699:{\displaystyle f(X)} 2681: 2643: 2586: 2548: 2513: 2384: 2364: 2325: 2289: 2262: 2223: 2200: 2154: 2127: 2121:statistical distance 2088: 2002: 1975: 1948: 1925: 1901: 1858: 1734: 1634: 1607: 1587: 1554: 1527: 1493: 1464: 1378: 1349: 1314: 1287: 1221:where the subscript 1149: 1119: 1099: 1057:{\displaystyle 1/n!} 1037: 953: 926: 905: 878: 834:, (−8, +8) for 717: 672: 601: 555: 526: 499: 463: 457:uncertainty analysis 439: 388: 348: 326: 259: 235: 203: 87:uncertainty analysis 74:Sensitivity analysis 5268:2016EnvMS..79..214P 5017:2005RiskA..25..481V 4937:1989Sci...246..221R 4902:2002JPCA..106.8721L 4859:2006JPCA..110.2474L 4815:2019GAMRE..13...53C 4718:2013MSSP...34...57B 4569:2019NatCo..10.4354W 4488:Sudret, B. (2008). 4411:2017EnvMS..95..115H 4366:2016WRR....52..440R 4317:2016WRR....52..423R 3819:10.1038/nature02771 3811:2004Natur.430..768M 3758:2005Sci...308...98B 3705:2008Sci...322.1085C 3699:(5904): 1085–1088. 3623:Statistical Science 3238:2008WatRM..22..393B 3090:Multiverse analysis 2967:), the White House 2961:European Commission 1941:, whence the name. 1651: 1239:Regression analysis 1234:Regression analysis 1003:partial derivatives 980:One-at-a-time (OAT) 752:partial derivatives 666:single-output codes 597:to approximate the 80:in the output of a 40:improve the content 5305:; Springer-Verlag. 4621:J. R. Stat. Soc. B 4278:. Elsevier: 1–11. 4253:10.1137/20M1354829 4218:10.3150/21-BEJ1438 4093:10.1007/BF02024507 4058:10.1111/rssb.12052 3286:10.1111/gwat.12330 2924:expert elicitation 2741:Gaussian processes 2696: 2667: 2625: 2572: 2534: 2414: 2370: 2346: 2295: 2268: 2241: 2206: 2182: 2140: 2109: 2074: 1988: 1971:, here denoted by 1961: 1931: 1907: 1871: 1844: 1720: 1637: 1620: 1593: 1580:total effect index 1567: 1540: 1510: 1479: 1450: 1362: 1320: 1300: 1208: 1132: 1105: 1093:partial derivative 1054: 966: 939: 911: 891: 872: 756:elementary effects 723: 678: 634:Correlated inputs: 607: 561: 532: 512: 469: 445: 425: 372: 332: 309: 241: 221:mathematical model 209: 197: 107:mathematical model 82:mathematical model 4931:(4927): 221–226. 4910:10.1021/jp014567t 4896:(37): 8721–8733. 4867:10.1021/jp054148m 4532:(11): 1735–1763. 4475:10.1137/130936233 3849:Czitrom, Veronica 3805:(7001): 768–772. 3671:978-1-61197-668-7 3561:(10): 4731–4744. 3485:10.1214/14-EJS895 3434:10.1214/12-EJS749 3383:10.1063/1.5120035 3168:Structural Safety 3133:978-0-470-05997-5 3095:Feature selection 3012:Model calibration 2965:impact assessment 2957:impact assessment 2796:Bayesian networks 2785:Smoothing splines 2769:Gradient boosting 2655: 2598: 2560: 2473:Chaos polynomials 2412: 2373:{\displaystyle Y} 2298:{\displaystyle i} 2209:{\displaystyle Y} 1934:{\displaystyle Y} 1910:{\displaystyle Y} 1718: 1596:{\displaystyle Y} 1448: 1339:main effect index 1323:{\displaystyle Y} 1243:linear regression 1194: 1182: 1108:{\displaystyle Y} 1007:linear regression 914:{\displaystyle Y} 726:{\displaystyle f} 711:data-driven model 681:{\displaystyle Y} 648:linear regression 610:{\displaystyle f} 591:data-driven model 564:{\displaystyle f} 535:{\displaystyle Y} 472:{\displaystyle Y} 448:{\displaystyle Y} 335:{\displaystyle Y} 244:{\displaystyle p} 212:{\displaystyle f} 68: 67: 60: 5376: 5291: 5289: 5279: 5229: 5228: 5222: 5214: 5212: 5211: 5205: 5199:. Archived from 5198: 5190: 5184: 5177: 5171: 5170: 5148: 5142: 5141: 5119: 5113: 5110: 5104: 5099: 5090: 5089: 5061: 5055: 5054: 5036: 4999: 4993: 4990: 4984: 4983: 4971: 4965: 4964: 4920: 4914: 4913: 4885: 4879: 4878: 4853:(7): 2474–2485. 4841: 4835: 4834: 4798: 4792: 4791: 4763: 4757: 4756: 4736: 4730: 4729: 4697: 4691: 4690: 4688: 4670: 4661: 4655: 4654: 4636: 4616: 4607: 4606: 4596: 4548: 4542: 4541: 4521: 4510: 4509: 4485: 4479: 4478: 4458: 4452: 4451: 4429: 4423: 4422: 4394: 4388: 4387: 4377: 4345: 4339: 4338: 4328: 4296: 4290: 4289: 4287: 4263: 4257: 4256: 4236: 4230: 4229: 4211: 4191: 4185: 4184: 4166: 4146: 4140: 4139: 4111: 4105: 4104: 4076: 4070: 4069: 4052:(5). : 925–947. 4041: 4035: 4034: 4022: 4010: 4004: 3998: 3992: 3991: 3983: 3977: 3976: 3968: 3962: 3961: 3953: 3947: 3946: 3918: 3912: 3911: 3909: 3883: 3877: 3876: 3845: 3839: 3838: 3792: 3786: 3785: 3752:(5718): 98–103. 3739: 3733: 3732: 3682: 3676: 3675: 3647: 3641: 3640: 3638: 3614: 3608: 3607: 3587: 3581: 3580: 3570: 3544: 3538: 3537: 3503: 3497: 3496: 3478: 3452: 3446: 3445: 3427: 3401: 3395: 3394: 3368: 3359: 3353: 3352: 3328: 3322: 3321: 3313: 3307: 3306: 3288: 3264: 3258: 3257: 3221: 3215: 3214: 3190: 3184: 3183: 3163: 3157: 3156: 3144: 3138: 3137: 3111: 2898:Edward E. Leamer 2839:Related concepts 2812:cross-validation 2804:machine learning 2705: 2703: 2702: 2697: 2676: 2674: 2673: 2668: 2657: 2656: 2648: 2634: 2632: 2631: 2626: 2600: 2599: 2591: 2581: 2579: 2578: 2573: 2562: 2561: 2553: 2543: 2541: 2540: 2535: 2503:machine learning 2423: 2421: 2420: 2415: 2413: 2411: 2410: 2409: 2396: 2388: 2379: 2377: 2376: 2371: 2355: 2353: 2352: 2347: 2304: 2302: 2301: 2296: 2277: 2275: 2274: 2269: 2250: 2248: 2247: 2242: 2215: 2213: 2212: 2207: 2191: 2189: 2188: 2183: 2181: 2180: 2179: 2178: 2169: 2149: 2147: 2146: 2141: 2139: 2138: 2118: 2116: 2115: 2110: 2083: 2081: 2080: 2075: 2067: 2066: 2065: 2064: 2055: 2039: 2038: 2014: 2013: 1997: 1995: 1994: 1989: 1987: 1986: 1970: 1968: 1967: 1962: 1960: 1959: 1940: 1938: 1937: 1932: 1916: 1914: 1913: 1908: 1880: 1878: 1877: 1872: 1870: 1869: 1853: 1851: 1850: 1845: 1840: 1839: 1815: 1814: 1796: 1795: 1765: 1764: 1749: 1748: 1729: 1727: 1726: 1721: 1719: 1717: 1703: 1696: 1695: 1674: 1662: 1650: 1645: 1629: 1627: 1626: 1621: 1619: 1618: 1602: 1600: 1599: 1594: 1576: 1574: 1573: 1568: 1566: 1565: 1549: 1547: 1546: 1541: 1539: 1538: 1519: 1517: 1516: 1511: 1500: 1488: 1486: 1485: 1480: 1459: 1457: 1456: 1451: 1449: 1447: 1433: 1426: 1425: 1407: 1395: 1390: 1389: 1371: 1369: 1368: 1363: 1361: 1360: 1329: 1327: 1326: 1321: 1309: 1307: 1306: 1301: 1299: 1298: 1270:random variables 1217: 1215: 1214: 1209: 1204: 1203: 1202: 1201: 1196: 1195: 1187: 1183: 1181: 1180: 1179: 1166: 1158: 1141: 1139: 1138: 1133: 1131: 1130: 1114: 1112: 1111: 1106: 1063: 1061: 1060: 1055: 1047: 975: 973: 972: 967: 965: 964: 948: 946: 945: 940: 938: 937: 920: 918: 917: 912: 900: 898: 897: 892: 890: 889: 732: 730: 729: 724: 693:Stochastic code: 687: 685: 684: 679: 616: 614: 613: 608: 570: 568: 567: 562: 541: 539: 538: 533: 521: 519: 518: 513: 511: 510: 478: 476: 475: 470: 454: 452: 451: 446: 434: 432: 431: 426: 400: 399: 381: 379: 378: 373: 341: 339: 338: 333: 318: 316: 315: 310: 305: 304: 280: 279: 250: 248: 247: 242: 227:"), viewed as a 225:programming code 218: 216: 215: 210: 63: 56: 52: 49: 43: 24: 23: 16: 5384: 5383: 5379: 5378: 5377: 5375: 5374: 5373: 5339: 5338: 5330: 5237: 5235:Further reading 5232: 5216: 5215: 5209: 5207: 5203: 5196: 5194:"Archived copy" 5192: 5191: 5187: 5178: 5174: 5149: 5145: 5120: 5116: 5111: 5107: 5100: 5093: 5062: 5058: 5000: 4996: 4991: 4987: 4972: 4968: 4921: 4917: 4886: 4882: 4842: 4838: 4799: 4795: 4764: 4760: 4737: 4733: 4698: 4694: 4668: 4662: 4658: 4617: 4610: 4549: 4545: 4522: 4513: 4486: 4482: 4459: 4455: 4448: 4430: 4426: 4395: 4391: 4346: 4342: 4297: 4293: 4264: 4260: 4237: 4233: 4192: 4188: 4147: 4143: 4112: 4108: 4077: 4073: 4042: 4038: 4011: 4007: 3999: 3995: 3984: 3980: 3969: 3965: 3956:Cacuci, Dan G. 3954: 3950: 3935:10.2307/1269043 3919: 3915: 3884: 3880: 3865:10.2307/2685731 3846: 3842: 3793: 3789: 3740: 3736: 3683: 3679: 3672: 3648: 3644: 3615: 3611: 3604: 3588: 3584: 3545: 3541: 3504: 3500: 3453: 3449: 3402: 3398: 3366: 3360: 3356: 3329: 3325: 3314: 3310: 3265: 3261: 3222: 3218: 3191: 3187: 3164: 3160: 3145: 3141: 3134: 3112: 3108: 3104: 3099: 3070:Robustification 3020: 2985: 2952: 2879: 2866: 2860: 2841: 2832: 2820: 2753:heteroscedastic 2743:(also known as 2682: 2679: 2678: 2647: 2646: 2644: 2641: 2640: 2590: 2589: 2587: 2584: 2583: 2552: 2551: 2549: 2546: 2545: 2514: 2511: 2510: 2495: 2483: 2475: 2462: 2460:Shapley effects 2449: 2443: 2405: 2401: 2397: 2389: 2387: 2385: 2382: 2381: 2365: 2362: 2361: 2326: 2323: 2322: 2319: 2290: 2287: 2286: 2263: 2260: 2259: 2224: 2221: 2220: 2201: 2198: 2197: 2174: 2170: 2165: 2161: 2157: 2155: 2152: 2151: 2134: 2130: 2128: 2125: 2124: 2089: 2086: 2085: 2060: 2056: 2051: 2047: 2043: 2034: 2030: 2009: 2005: 2003: 2000: 1999: 1982: 1978: 1976: 1973: 1972: 1955: 1951: 1949: 1946: 1945: 1926: 1923: 1922: 1902: 1899: 1898: 1895: 1865: 1861: 1859: 1856: 1855: 1835: 1831: 1804: 1800: 1785: 1781: 1760: 1756: 1741: 1737: 1735: 1732: 1731: 1704: 1688: 1684: 1670: 1663: 1661: 1646: 1641: 1635: 1632: 1631: 1614: 1610: 1608: 1605: 1604: 1588: 1585: 1584: 1561: 1557: 1555: 1552: 1551: 1534: 1530: 1528: 1525: 1524: 1496: 1494: 1491: 1490: 1465: 1462: 1461: 1434: 1421: 1417: 1403: 1396: 1394: 1385: 1381: 1379: 1376: 1375: 1356: 1352: 1350: 1347: 1346: 1315: 1312: 1311: 1294: 1290: 1288: 1285: 1284: 1266: 1260: 1236: 1197: 1191: 1190: 1189: 1188: 1175: 1171: 1167: 1159: 1157: 1153: 1152: 1150: 1147: 1146: 1126: 1122: 1120: 1117: 1116: 1100: 1097: 1096: 1095:of the output 1089: 1080: 1074: 1043: 1038: 1035: 1034: 1031:hyperoctahedron 988: 982: 960: 956: 954: 951: 950: 933: 929: 927: 924: 923: 906: 903: 902: 901:and the output 885: 881: 879: 876: 875: 861: 854: 847: 840: 833: 826: 819: 812: 805: 789: 787:Visual analysis 743: 718: 715: 714: 673: 670: 669: 602: 599: 598: 556: 553: 552: 548: 527: 524: 523: 506: 502: 500: 497: 496: 479:(providing its 464: 461: 460: 440: 437: 436: 395: 391: 389: 386: 385: 349: 346: 345: 327: 324: 323: 300: 296: 275: 271: 260: 257: 256: 236: 233: 232: 204: 201: 200: 180: 103: 71: 64: 53: 47: 44: 37: 33:Manual of Style 25: 21: 12: 11: 5: 5382: 5372: 5371: 5366: 5361: 5356: 5351: 5337: 5336: 5329: 5328:External links 5326: 5325: 5324: 5317: 5306: 5299: 5292: 5247: 5236: 5233: 5231: 5230: 5185: 5172: 5161:(3): 308–313. 5143: 5114: 5105: 5091: 5056: 5011:(2): 481–492. 4994: 4985: 4966: 4915: 4880: 4836: 4793: 4774:(4): 367–388. 4758: 4747:(7): 964–979. 4731: 4712:(1–2): 57–75. 4692: 4656: 4627:(3): 751–769. 4608: 4543: 4511: 4500:(7): 964–979. 4480: 4453: 4446: 4424: 4389: 4360:(1): 440–455. 4340: 4311:(1): 423–439. 4291: 4258: 4231: 4186: 4141: 4106: 4087:(3): 441–451. 4071: 4036: 4017:(in Russian). 4005: 3993: 3978: 3963: 3948: 3913: 3900:(1): 125–153. 3878: 3859:(2): 126–131. 3840: 3787: 3734: 3677: 3670: 3642: 3629:(4): 409–435. 3609: 3602: 3582: 3539: 3520:(3): 833–847. 3498: 3447: 3396: 3354: 3323: 3308: 3279:(2): 159–170. 3259: 3232:(3): 293–408. 3216: 3205:(2): 139–152. 3185: 3174:(2): 105–112. 3158: 3139: 3132: 3105: 3103: 3100: 3098: 3097: 3092: 3087: 3082: 3077: 3072: 3067: 3062: 3057: 3052: 3047: 3042: 3037: 3032: 3027: 3021: 3019: 3016: 3015: 3014: 3009: 3004: 2999: 2994: 2984: 2981: 2951: 2948: 2947: 2946: 2940: 2934: 2927: 2916: 2915: 2902: 2901: 2878: 2875: 2862:Main article: 2859: 2856: 2840: 2837: 2831: 2828: 2819: 2816: 2800: 2799: 2792: 2782: 2772: 2766: 2763:decision trees 2759:Random forests 2756: 2726:Sobol sequence 2724:, such as the 2718: 2717: 2714: 2711: 2695: 2692: 2689: 2686: 2666: 2663: 2660: 2654: 2651: 2624: 2621: 2618: 2615: 2612: 2609: 2606: 2603: 2597: 2594: 2571: 2568: 2565: 2559: 2556: 2533: 2530: 2527: 2524: 2521: 2518: 2494: 2491: 2482: 2479: 2474: 2471: 2466:Shapley values 2461: 2458: 2453:Fourier series 2445:Main article: 2442: 2439: 2435:variance-based 2408: 2404: 2400: 2395: 2392: 2369: 2345: 2342: 2339: 2336: 2333: 2330: 2318: 2311: 2294: 2267: 2240: 2237: 2234: 2231: 2228: 2205: 2177: 2173: 2168: 2164: 2160: 2137: 2133: 2108: 2105: 2102: 2099: 2096: 2093: 2073: 2070: 2063: 2059: 2054: 2050: 2046: 2042: 2037: 2033: 2029: 2026: 2023: 2020: 2017: 2012: 2008: 1985: 1981: 1958: 1954: 1930: 1906: 1894: 1891: 1868: 1864: 1843: 1838: 1834: 1830: 1827: 1824: 1821: 1818: 1813: 1810: 1807: 1803: 1799: 1794: 1791: 1788: 1784: 1780: 1777: 1774: 1771: 1768: 1763: 1759: 1755: 1752: 1747: 1744: 1740: 1716: 1713: 1710: 1707: 1702: 1699: 1694: 1691: 1687: 1683: 1680: 1677: 1673: 1669: 1666: 1660: 1657: 1654: 1649: 1644: 1640: 1617: 1613: 1592: 1564: 1560: 1537: 1533: 1509: 1506: 1503: 1499: 1478: 1475: 1472: 1469: 1446: 1443: 1440: 1437: 1432: 1429: 1424: 1420: 1416: 1413: 1410: 1406: 1402: 1399: 1393: 1388: 1384: 1359: 1355: 1319: 1297: 1293: 1262:Main article: 1259: 1256: 1235: 1232: 1219: 1218: 1207: 1200: 1186: 1178: 1174: 1170: 1165: 1162: 1156: 1129: 1125: 1104: 1088: 1085: 1076:Main article: 1073: 1070: 1053: 1050: 1046: 1042: 999: 998: 995: 984:Main article: 981: 978: 963: 959: 936: 932: 910: 888: 884: 859: 852: 845: 838: 831: 824: 817: 810: 803: 788: 785: 777: 776: 773: 766: 763: 742: 739: 735: 734: 722: 698: 697: 690: 677: 659: 641: 631: 630: 629: 618: 606: 560: 547: 544: 531: 509: 505: 468: 444: 424: 421: 418: 415: 412: 409: 406: 403: 398: 394: 371: 368: 365: 362: 359: 356: 353: 331: 308: 303: 299: 295: 292: 289: 286: 283: 278: 274: 270: 267: 264: 240: 208: 179: 176: 175: 174: 171: 168: 157: 154: 151: 148: 145: 142: 102: 99: 69: 66: 65: 48:September 2024 28: 26: 19: 9: 6: 4: 3: 2: 5381: 5370: 5367: 5365: 5362: 5360: 5357: 5355: 5352: 5350: 5347: 5346: 5344: 5335: 5332: 5331: 5322: 5318: 5315: 5311: 5310:Komkov, Vadim 5307: 5304: 5300: 5297: 5293: 5288: 5283: 5278: 5273: 5269: 5265: 5261: 5257: 5253: 5248: 5246: 5243: 5239: 5238: 5226: 5220: 5206:on 2011-04-26 5202: 5195: 5189: 5182: 5176: 5168: 5164: 5160: 5156: 5155: 5147: 5139: 5135: 5131: 5127: 5126: 5118: 5109: 5103: 5098: 5096: 5087: 5083: 5079: 5075: 5071: 5067: 5060: 5052: 5048: 5044: 5040: 5035: 5030: 5026: 5022: 5018: 5014: 5010: 5006: 5005:Risk Analysis 4998: 4989: 4981: 4977: 4970: 4962: 4958: 4954: 4950: 4946: 4942: 4938: 4934: 4930: 4926: 4919: 4911: 4907: 4903: 4899: 4895: 4891: 4884: 4876: 4872: 4868: 4864: 4860: 4856: 4852: 4848: 4840: 4832: 4828: 4824: 4820: 4816: 4812: 4808: 4804: 4797: 4789: 4785: 4781: 4777: 4773: 4769: 4762: 4754: 4750: 4746: 4742: 4735: 4727: 4723: 4719: 4715: 4711: 4707: 4703: 4696: 4687: 4682: 4678: 4674: 4667: 4660: 4652: 4648: 4644: 4640: 4635: 4634:10.1.1.6.9720 4630: 4626: 4622: 4615: 4613: 4604: 4600: 4595: 4590: 4586: 4582: 4578: 4574: 4570: 4566: 4562: 4558: 4554: 4547: 4539: 4535: 4531: 4527: 4520: 4518: 4516: 4507: 4503: 4499: 4495: 4491: 4484: 4476: 4472: 4468: 4464: 4457: 4449: 4447:9780128030318 4443: 4439: 4435: 4428: 4420: 4416: 4412: 4408: 4404: 4400: 4393: 4385: 4381: 4376: 4371: 4367: 4363: 4359: 4355: 4351: 4344: 4336: 4332: 4327: 4322: 4318: 4314: 4310: 4306: 4302: 4295: 4286: 4281: 4277: 4273: 4269: 4262: 4254: 4250: 4246: 4242: 4235: 4227: 4223: 4219: 4215: 4210: 4205: 4201: 4197: 4190: 4182: 4178: 4174: 4170: 4165: 4160: 4156: 4152: 4145: 4137: 4133: 4129: 4125: 4121: 4117: 4110: 4102: 4098: 4094: 4090: 4086: 4082: 4075: 4067: 4063: 4059: 4055: 4051: 4047: 4040: 4032: 4028: 4020: 4016: 4009: 4003: 3997: 3989: 3982: 3974: 3967: 3959: 3952: 3944: 3940: 3936: 3932: 3928: 3924: 3923:Technometrics 3917: 3908: 3903: 3899: 3895: 3894: 3889: 3882: 3874: 3870: 3866: 3862: 3858: 3854: 3850: 3844: 3836: 3832: 3828: 3824: 3820: 3816: 3812: 3808: 3804: 3800: 3799: 3791: 3783: 3779: 3775: 3771: 3767: 3763: 3759: 3755: 3751: 3747: 3746: 3738: 3730: 3726: 3722: 3718: 3714: 3710: 3706: 3702: 3698: 3694: 3693: 3688: 3681: 3673: 3667: 3663: 3659: 3655: 3654: 3646: 3637: 3632: 3628: 3624: 3620: 3613: 3605: 3603:9780470033302 3599: 3595: 3594: 3586: 3578: 3574: 3569: 3564: 3560: 3556: 3555: 3550: 3543: 3535: 3531: 3527: 3523: 3519: 3515: 3514: 3509: 3502: 3494: 3490: 3486: 3482: 3477: 3472: 3468: 3464: 3463: 3458: 3451: 3443: 3439: 3435: 3431: 3426: 3421: 3418:: 2420–2448. 3417: 3413: 3412: 3407: 3400: 3392: 3388: 3384: 3380: 3377:(5): 053303. 3376: 3372: 3365: 3358: 3350: 3346: 3342: 3338: 3334: 3327: 3319: 3312: 3304: 3300: 3296: 3292: 3287: 3282: 3278: 3274: 3270: 3263: 3255: 3251: 3247: 3243: 3239: 3235: 3231: 3227: 3220: 3212: 3208: 3204: 3200: 3196: 3189: 3181: 3177: 3173: 3169: 3162: 3154: 3150: 3143: 3135: 3129: 3125: 3121: 3117: 3110: 3106: 3096: 3093: 3091: 3088: 3086: 3083: 3081: 3078: 3076: 3073: 3071: 3068: 3066: 3063: 3061: 3058: 3056: 3053: 3051: 3048: 3046: 3043: 3041: 3038: 3036: 3033: 3031: 3028: 3026: 3023: 3022: 3013: 3010: 3008: 3005: 3003: 3000: 2998: 2995: 2993: 2990: 2989: 2988: 2980: 2978: 2974: 2970: 2966: 2962: 2958: 2944: 2941: 2938: 2935: 2931: 2928: 2925: 2921: 2918: 2917: 2914: 2909: 2908: 2906: 2899: 2895: 2891: 2887: 2884: 2883: 2882: 2874: 2870: 2865: 2855: 2853: 2848: 2846: 2836: 2827: 2825: 2815: 2813: 2809: 2805: 2797: 2793: 2790: 2786: 2783: 2780: 2776: 2773: 2770: 2767: 2764: 2760: 2757: 2754: 2750: 2746: 2742: 2739: 2738: 2737: 2735: 2731: 2730:Ilya M. Sobol 2727: 2723: 2715: 2712: 2709: 2708: 2707: 2690: 2684: 2661: 2649: 2637: 2619: 2613: 2610: 2604: 2592: 2566: 2554: 2528: 2522: 2519: 2516: 2508: 2504: 2500: 2499:data-modeling 2490: 2488: 2478: 2470: 2467: 2457: 2454: 2448: 2438: 2436: 2431: 2425: 2406: 2402: 2393: 2367: 2359: 2340: 2334: 2331: 2328: 2316: 2310: 2308: 2292: 2284: 2279: 2265: 2256: 2254: 2238: 2235: 2226: 2217: 2203: 2195: 2175: 2171: 2162: 2158: 2135: 2131: 2122: 2103: 2100: 2097: 2091: 2061: 2057: 2048: 2044: 2040: 2035: 2031: 2024: 2018: 2015: 2010: 2006: 1983: 1979: 1956: 1952: 1942: 1928: 1920: 1904: 1890: 1888: 1882: 1866: 1862: 1836: 1832: 1828: 1825: 1822: 1819: 1816: 1811: 1808: 1805: 1801: 1797: 1792: 1789: 1786: 1782: 1778: 1775: 1772: 1769: 1766: 1761: 1757: 1750: 1745: 1742: 1738: 1711: 1705: 1692: 1689: 1685: 1678: 1664: 1658: 1655: 1652: 1647: 1642: 1638: 1615: 1611: 1590: 1582: 1581: 1562: 1558: 1535: 1531: 1521: 1504: 1473: 1467: 1441: 1435: 1422: 1418: 1411: 1397: 1391: 1386: 1382: 1373: 1357: 1353: 1345:For an input 1343: 1341: 1340: 1335: 1334: 1317: 1295: 1291: 1282: 1281:Sobol indices 1277: 1275: 1271: 1265: 1255: 1253: 1248: 1244: 1240: 1231: 1227: 1224: 1205: 1198: 1184: 1176: 1172: 1163: 1154: 1145: 1144: 1143: 1127: 1123: 1102: 1094: 1084: 1079: 1078:Morris_method 1069: 1067: 1066:linear models 1051: 1048: 1044: 1040: 1032: 1028: 1024: 1018: 1016: 1010: 1008: 1004: 996: 993: 992: 991: 987: 977: 961: 957: 934: 930: 908: 886: 882: 869: 865: 858: 851: 844: 837: 830: 823: 816: 809: 802: 798: 793: 784: 780: 774: 771: 767: 764: 761: 760: 759: 757: 753: 749: 738: 720: 712: 708: 703: 700: 699: 694: 691: 675: 667: 663: 660: 657: 653: 649: 645: 644:Nonlinearity: 642: 639: 635: 632: 627: 623: 619: 604: 596: 592: 588: 584: 583: 581: 577: 574: 573: 572: 558: 543: 529: 507: 503: 494: 490: 486: 482: 466: 458: 442: 422: 419: 416: 413: 410: 407: 404: 401: 396: 392: 382: 369: 363: 357: 354: 351: 343: 329: 322: 301: 297: 293: 290: 287: 284: 281: 276: 272: 265: 262: 254: 251:-dimensional 238: 230: 226: 222: 206: 194: 190: 184: 172: 169: 166: 162: 158: 155: 152: 149: 146: 143: 140: 136: 135: 134: 130: 128: 124: 120: 116: 112: 108: 98: 96: 92: 88: 83: 79: 75: 62: 59: 51: 41: 36: 34: 27: 18: 17: 5320: 5313: 5302: 5295: 5259: 5255: 5241: 5208:. Retrieved 5201:the original 5188: 5180: 5175: 5158: 5152: 5146: 5132:(1): 31–43. 5129: 5123: 5117: 5108: 5069: 5065: 5059: 5008: 5004: 4997: 4988: 4979: 4975: 4969: 4928: 4924: 4918: 4893: 4889: 4883: 4850: 4846: 4839: 4809:(1): 53–65. 4806: 4802: 4796: 4771: 4767: 4761: 4744: 4740: 4734: 4709: 4705: 4695: 4676: 4672: 4659: 4624: 4620: 4560: 4556: 4546: 4529: 4525: 4497: 4493: 4483: 4466: 4462: 4456: 4437: 4427: 4402: 4398: 4392: 4357: 4353: 4343: 4308: 4304: 4294: 4275: 4271: 4261: 4244: 4240: 4234: 4199: 4195: 4189: 4154: 4150: 4144: 4119: 4115: 4109: 4084: 4080: 4074: 4049: 4045: 4039: 4030: 4026: 4018: 4014: 4008: 3996: 3987: 3981: 3972: 3966: 3957: 3951: 3926: 3922: 3916: 3897: 3891: 3881: 3856: 3852: 3843: 3802: 3796: 3790: 3749: 3743: 3737: 3696: 3690: 3680: 3652: 3645: 3626: 3622: 3612: 3592: 3585: 3558: 3552: 3542: 3517: 3511: 3501: 3466: 3460: 3450: 3415: 3409: 3399: 3374: 3370: 3357: 3340: 3336: 3326: 3317: 3311: 3276: 3272: 3262: 3229: 3225: 3219: 3202: 3198: 3188: 3171: 3167: 3161: 3152: 3148: 3142: 3115: 3109: 3050:Interval FEM 3002:Epidemiology 2986: 2953: 2942: 2936: 2929: 2919: 2912: 2904: 2885: 2880: 2871: 2867: 2849: 2842: 2833: 2821: 2801: 2777:, which use 2719: 2638: 2582:, such that 2506: 2496: 2484: 2476: 2463: 2450: 2426: 2320: 2314: 2280: 2257: 2218: 2196:measures of 1943: 1896: 1883: 1579: 1578: 1522: 1374: 1344: 1338: 1337: 1332: 1331: 1280: 1278: 1267: 1237: 1228: 1222: 1220: 1090: 1081: 1019: 1015:interactions 1011: 1000: 989: 873: 867: 863: 856: 849: 842: 835: 828: 821: 814: 807: 800: 796: 781: 778: 744: 736: 701: 692: 661: 643: 638:independence 633: 593:) including 575: 549: 383: 344: 320: 252: 224: 220: 198: 161:optimization 137:Testing the 131: 117:, including 104: 73: 72: 54: 45: 38:Please help 30: 5287:10871/21086 5262:: 214–232. 5034:1874/386039 4563:(1): 4354. 4405:: 115–131. 3469:: 575–603. 3273:Groundwater 2890:assumptions 2845:uncertainty 1887:Monte Carlo 1023:convex hull 733:-function). 231:, with the 219:, (called " 123:reliability 115:uncertainty 78:uncertainty 5354:Simulation 5343:Categories 5210:2009-10-16 5072:: 102436. 4209:2102.00356 4164:1909.10140 4033:: 407–414. 4021:: 112–118. 3102:References 2507:metamodels 2493:Metamodels 2430:variograms 2358:variograms 1603:caused by 1027:octahedron 707:meta-model 689:interpret. 587:meta-model 481:statistics 193:pie charts 139:robustness 127:stochastic 101:Motivation 5086:198636712 4831:216590427 4629:CiteSeerX 4585:2041-1723 4384:1944-7973 4335:1944-7973 4226:1350-7265 4196:Bernoulli 4181:0162-1459 4136:0951-8320 4101:1588-2632 4066:1369-7412 3729:206515456 3568:0802.1099 3534:0960-3174 3493:1935-7524 3476:1311.1797 3442:1935-7524 3425:1112.1788 3391:208835771 3075:ROC curve 3025:Causality 2894:inference 2808:nonlinear 2794:Discrete 2653:^ 2611:≈ 2596:^ 2558:^ 2399:∂ 2391:∂ 2266:δ 2236:≥ 2104:⋅ 2098:⋅ 2007:ξ 1980:ξ 1790:− 1743:∼ 1690:∼ 1659:− 1505:⋅ 1474:⋅ 1169:∂ 1161:∂ 652:nonlinear 417:… 229:black box 111:black box 5219:cite web 5051:15988654 5043:15876219 4961:23088466 4953:17839016 4875:16480307 4603:31554788 3827:15306806 3782:14404609 3774:15802601 3721:19008442 3656:. SIAM. 3303:25810333 3018:See also 2997:Business 2253:distance 2084:, where 580:sampling 455:. While 319:and the 129:events. 5312:(1986) 5264:Bibcode 5167:1814801 5138:1803924 5066:Futures 5013:Bibcode 4982:: 7–18. 4933:Bibcode 4925:Science 4898:Bibcode 4855:Bibcode 4811:Bibcode 4788:7678955 4714:Bibcode 4651:6130150 4594:6761138 4565:Bibcode 4407:Bibcode 4362:Bibcode 4313:Bibcode 3990:. SIAM. 3943:1269043 3873:2685731 3807:Bibcode 3754:Bibcode 3745:Science 3701:Bibcode 3692:Science 3295:1286771 3254:9710579 3234:Bibcode 2745:kriging 485:moments 255:vector 5165: 5136: 5084: 5049: 5041: 4959: 4951: 4873: 4829: 4786: 4649: 4631: 4601: 4591: 4583: 4444: 4382: 4333: 4224: 4179: 4134: 4099: 4064: 3941: 3871: 3835:980153 3833: 3825: 3798:Nature 3780: 3772: 3727: 3719: 3668: 3600: 3532: 3491: 3440: 3389: 3301: 3293: 3252: 3130: 2971:, the 1919:moment 1730:where 1460:where 1072:Morris 321:output 223:" ou " 5204:(PDF) 5197:(PDF) 5163:JSTOR 5134:JSTOR 5082:S2CID 5047:S2CID 4957:S2CID 4827:S2CID 4784:S2CID 4679:(6). 4669:(PDF) 4647:S2CID 4204:arXiv 4159:arXiv 4002:IEEE. 3939:JSTOR 3869:JSTOR 3831:S2CID 3778:S2CID 3725:S2CID 3563:arXiv 3471:arXiv 3420:arXiv 3387:S2CID 3367:(PDF) 3250:S2CID 2251:is a 2119:is a 253:input 5225:link 5039:PMID 4949:PMID 4871:PMID 4599:PMID 4581:ISSN 4442:ISBN 4380:ISSN 4331:ISSN 4222:ISSN 4177:ISSN 4132:ISSN 4097:ISSN 4062:ISSN 3823:PMID 3770:PMID 3717:PMID 3666:ISBN 3598:ISBN 3530:ISSN 3489:ISSN 3438:ISSN 3299:PMID 3291:OSTI 3128:ISBN 2975:and 2315:VARS 2150:and 1921:of 1881:. 1489:and 949:and 848:and 595:HDMR 163:and 93:and 5282:hdl 5272:doi 5074:doi 5070:112 5029:hdl 5021:doi 4941:doi 4929:246 4906:doi 4894:106 4863:doi 4851:110 4819:doi 4776:doi 4749:doi 4722:doi 4681:doi 4639:doi 4589:PMC 4573:doi 4534:doi 4502:doi 4471:doi 4415:doi 4370:doi 4321:doi 4280:doi 4249:doi 4214:doi 4169:doi 4155:116 4124:doi 4089:doi 4054:doi 3931:doi 3902:doi 3898:169 3861:doi 3815:doi 3803:430 3762:doi 3750:308 3709:doi 3697:322 3658:doi 3631:doi 3573:doi 3522:doi 3481:doi 3430:doi 3379:doi 3345:doi 3281:doi 3242:doi 3207:doi 3176:doi 3120:doi 2732:or 2219:If 1336:or 1005:or 754:or 493:cdf 489:pdf 189:pdf 5345:: 5280:. 5270:. 5260:79 5258:. 5254:. 5221:}} 5217:{{ 5159:75 5157:. 5130:73 5128:. 5094:^ 5080:. 5068:. 5045:. 5037:. 5027:. 5019:. 5009:25 5007:. 4978:. 4955:. 4947:. 4939:. 4927:. 4904:. 4892:. 4869:. 4861:. 4849:. 4825:. 4817:. 4807:13 4805:. 4782:. 4772:94 4770:. 4745:93 4743:. 4720:. 4710:34 4708:. 4704:. 4677:33 4675:. 4671:. 4645:. 4637:. 4625:66 4623:. 4611:^ 4597:. 4587:. 4579:. 4571:. 4561:10 4559:. 4555:. 4530:94 4528:. 4514:^ 4498:93 4496:. 4492:. 4465:. 4413:. 4403:95 4401:. 4378:. 4368:. 4358:52 4356:. 4352:. 4329:. 4319:. 4309:52 4307:. 4303:. 4276:67 4274:. 4270:. 4245:10 4243:. 4220:. 4212:. 4200:28 4198:. 4175:. 4167:. 4153:. 4130:. 4120:92 4118:. 4095:. 4085:10 4083:. 4060:. 4050:76 4048:. 4029:. 3937:. 3927:33 3925:. 3896:. 3890:. 3867:. 3857:53 3855:. 3829:. 3821:. 3813:. 3801:. 3776:. 3768:. 3760:. 3748:. 3723:. 3715:. 3707:. 3695:. 3689:. 3664:. 3625:. 3621:. 3571:. 3559:52 3557:. 3551:. 3528:. 3518:22 3516:. 3510:. 3487:. 3479:. 3465:. 3459:. 3436:. 3428:. 3414:. 3408:. 3385:. 3375:11 3373:. 3369:. 3341:91 3339:. 3335:. 3297:. 3289:. 3277:54 3275:. 3271:. 3248:. 3240:. 3230:22 3228:. 3203:16 3201:. 3197:. 3172:31 3170:. 3151:. 3126:. 3118:. 2822:A 2814:. 2424:. 2216:. 1342:. 1142:: 855:. 820:, 813:, 806:, 750:, 709:, 589:, 491:, 487:, 483:, 167:). 105:A 5290:. 5284:: 5274:: 5266:: 5227:) 5213:. 5169:. 5140:. 5088:. 5076:: 5053:. 5031:: 5023:: 5015:: 4980:7 4963:. 4943:: 4935:: 4912:. 4908:: 4900:: 4877:. 4865:: 4857:: 4833:. 4821:: 4813:: 4790:. 4778:: 4755:. 4751:: 4728:. 4724:: 4716:: 4689:. 4683:: 4653:. 4641:: 4605:. 4575:: 4567:: 4540:. 4536:: 4508:. 4504:: 4477:. 4473:: 4467:2 4450:. 4421:. 4417:: 4409:: 4386:. 4372:: 4364:: 4337:. 4323:: 4315:: 4288:. 4282:: 4255:. 4251:: 4228:. 4216:: 4206:: 4183:. 4171:: 4161:: 4138:. 4126:: 4103:. 4091:: 4068:. 4056:: 4031:1 4019:2 3945:. 3933:: 3910:. 3904:: 3875:. 3863:: 3837:. 3817:: 3809:: 3784:. 3764:: 3756:: 3731:. 3711:: 3703:: 3674:. 3660:: 3639:. 3633:: 3627:4 3606:. 3579:. 3575:: 3565:: 3536:. 3524:: 3495:. 3483:: 3473:: 3467:8 3444:. 3432:: 3422:: 3416:6 3393:. 3381:: 3351:. 3347:: 3305:. 3283:: 3256:. 3244:: 3236:: 3213:. 3209:: 3182:. 3178:: 3155:. 3153:1 3136:. 3122:: 2900:: 2694:) 2691:X 2688:( 2685:f 2665:) 2662:X 2659:( 2650:f 2623:) 2620:X 2617:( 2614:f 2608:) 2605:X 2602:( 2593:f 2570:) 2567:X 2564:( 2555:f 2532:) 2529:X 2526:( 2523:f 2520:= 2517:Y 2501:/ 2407:i 2403:x 2394:Y 2368:Y 2344:) 2341:X 2338:( 2335:f 2332:= 2329:Y 2317:) 2293:i 2239:0 2233:) 2230:( 2227:d 2204:Y 2176:i 2172:X 2167:| 2163:Y 2159:P 2136:Y 2132:P 2107:) 2101:, 2095:( 2092:d 2072:] 2069:) 2062:i 2058:X 2053:| 2049:Y 2045:P 2041:, 2036:Y 2032:P 2028:( 2025:d 2022:[ 2019:E 2016:= 2011:i 1984:i 1957:i 1953:X 1929:Y 1905:Y 1867:i 1863:X 1842:) 1837:p 1833:X 1829:, 1826:. 1823:. 1820:. 1817:, 1812:1 1809:+ 1806:i 1802:X 1798:, 1793:1 1787:i 1783:X 1779:, 1776:. 1773:. 1770:. 1767:, 1762:1 1758:X 1754:( 1751:= 1746:i 1739:X 1715:) 1712:Y 1709:( 1706:V 1701:) 1698:] 1693:i 1686:X 1682:| 1679:Y 1676:[ 1672:E 1668:( 1665:V 1656:1 1653:= 1648:T 1643:i 1639:S 1616:i 1612:X 1591:Y 1563:i 1559:X 1536:i 1532:X 1508:] 1502:[ 1498:E 1477:) 1471:( 1468:V 1445:) 1442:Y 1439:( 1436:V 1431:) 1428:] 1423:i 1419:X 1415:| 1412:Y 1409:[ 1405:E 1401:( 1398:V 1392:= 1387:i 1383:S 1358:i 1354:X 1318:Y 1296:i 1292:X 1223:x 1206:, 1199:0 1193:x 1185:| 1177:i 1173:X 1164:Y 1155:| 1128:i 1124:X 1103:Y 1052:! 1049:n 1045:/ 1041:1 962:4 958:Z 935:3 931:Z 909:Y 887:i 883:Z 870:. 868:Y 864:Y 860:4 857:Z 853:4 850:Z 846:3 843:Z 839:2 836:Z 832:1 829:Z 825:4 822:Z 818:3 815:Z 811:2 808:Z 804:1 801:Z 797:Y 721:f 676:Y 605:f 559:f 530:Y 508:i 504:X 467:Y 443:Y 423:p 420:, 414:, 411:1 408:= 405:i 402:, 397:i 393:X 370:. 367:) 364:X 361:( 358:f 355:= 352:Y 330:Y 307:) 302:p 298:X 294:, 291:. 288:. 285:. 282:, 277:1 273:X 269:( 266:= 263:X 239:p 207:f 61:) 55:( 50:) 46:( 42:. 35:.

Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.

Index