China 
Africa 
Explore chapters and articles related to this topic
Machine learning approach for stress analyses of steel members affected by elastic shear lag
Published in Alphose Zingoni, Current Perspectives and New Directions in Mechanics, Modelling and Design of Structural Systems, 2022
In the context of interpretable ML models, influences of input parameters on the output are examined through a variance-based sensitivity analysis using the “total-effect” index (see Saltelli (2002) for details). The total-effect index of the input parameters can be interpreted as the contribution to the output variance, including the impact of all interactions with other input variables. The results demonstrate that the influences of the top flange btf and the system length L increase towards the edges of the flange (y = ± btf/2), while the influences of the top flange thickness and the web height decrease, as displayed in Figure 4.
Global sensitivity analysis on thermo-hydro-mechanical coupled processes in a low strength sedimentary rock
Published in Vladimir Litvinenko, EUROCK2018: Geomechanics and Geodynamics of Rock Masses, 2018
Samuel Parsons, Graham Stuart, Bill Murphy, David Price
The three sensitivity analyses are the Regional Sensitivity Analysis, PAWN (Pianosi et al., 2015), and an estimate of the main effects indices from the Variance-based Sensitivity Analysis (Petropoulos and Srivastava, 2016). SAFE Toolbox (Pianosi et al., 2015) contains Matlab functions for calculating the sensitivity indices of these approaches using X and Y.
Probabilistic modeling and prediction of out-of-plane unidirectional composite lamina properties
Published in Mechanics of Advanced Materials and Structures, 2021
Jiaxin Zhang, Michael Shields, Stephanie TerMaath
Variance-based sensitivity analysis decomposes the variance of the output of the model or system into fractions that can be attributed to each input parameter. A specific measure of variance-based sensitivity that is commonly used is the so-called Sobol indices [74] that are defined as the relative contribution of the partial variances to the total variance V such that: where Si is commonly referred to as the first-order index that measures the contribution of each input variable xi to the variance of model output y taken separately without interacting with any other inputs.Sij is the second-order index which estimates the contribution of interactions between variable xi and xj to the total variance. One can infer the impact of each input variable and the interaction of variables on the output variance using Si, Sij and high-order indices in Eq. (20). Another popular variance based measure known as the total-order index is typically used to estimate the contribution of variable xi and its interactions to the output variance. It is given as follows:
An empirical analysis of intention of use for bike-sharing system in China through machine learning techniques
Published in Enterprise Information Systems, 2021
Tao Zhou, Kris M. Y. Law, K. L. Yung
Sensitivity analysis is a means of identifying how the uncertainty of output variable can be influenced by its dependent input variables (Leong et al. 2015). Through conducting a sensitivity analysis, the relative importance of inputs on the output can be quantified. One-factor-at-a-time algorithm is one of the most widely used sensitivity methods which analyses the sensitivity by changing the values of one variable at a time while keeping the other variables fixed (Jung et al. 2010). However, it is not applicable to nonlinear models. Therefore, in this section, Sobol’s sensitivity analysis is adopted. As a popular method of variance based sensitivity analysis, Sobol’s sensitivity analysis aims to measure the variance of the contribution to the output that each input variable has made, which is capable of measuring the sensitivity across the whole input space and dealing with nonlinear relationships (Rosolem et al. 2012).
Global sensitivity analysis of soil structure interaction system using N2-SSI method
Published in European Journal of Environmental and Civil Engineering, 2018
M. Zoutat, S.M. Elachachi, M. Mekki, M. Hamane
This method was developed by Sobol (1993). It is a variance-based sensitivity analysis technique, based on total sensitivity indices that take into account interaction effects between inputs. Sobol’s method is implemented by decomposing the input–output relationship into a summand of increasing dimensionality. Solving the resulting equation to obtain sensitivity indices requires the use of numerical integration techniques. The contribution of each parameter to the variance of the output amounts to study how the variance V(y) decreases if the xi variable is set to value . It is to calculate its expectation for all possible values of . Considering the expression of the total variance given by: