Automatic selection for general surrogate models
read more
Citations
Inverse modeling of saturated-unsaturated flow in site-scale fractured rocks using the continuum approach: A case study at Baihetan dam site, Southwest China
Modelling for Digital Twins—Potential Role of Surrogate Models
Low-fidelity scale factor improves Bayesian multi-fidelity prediction by reducing bumpiness of discrepancy function
Efficient global optimization with ensemble and selection of kernel functions for engineering design
Understanding the Effect of Hyperparameter Optimization on Machine Learning Models for Structure Design Problems
References
A new look at the statistical model identification
Estimating the Dimension of a Model
An introduction to the bootstrap
Estimating the dimension of a model
Related Papers (5)
Structural reliability analysis based on ensemble learning of surrogate models
The design and analysis of computer experiments
Frequently Asked Questions (8)
Q2. What are the contributions mentioned in the paper "Automatic selection for general surrogate models" ?
In this paper, the authors introduce a universal criterion that can be applied to any type of surrogate models. Based on this criterion, the authors propose two automatic selection algorithms. The results show the efficiency of their approach. The second selection scheme further explores the space of surrogate models by using an evolutionary algorithm where each individual is a surrogate model.
Q3. What is the evolution of surrogate models?
The mutation and cross-over operators between two surrogate models of the same type are performed by modifying or exchanging the surrogate models settings.
Q4. What is the optimal weighted surrogate OWS?
The optimal weighted surrogate OWS is obtained using the weights of Equation (8).W = C−111>C−11 (8)where the elements of the matrix C, ci j =< vi,vj >. (Viana et al, 2009) noticed that the solution may include negative values.
Q5. What is the way to select a surrogate model?
For instance, to tune a universal kriging surrogate model, there are various possible choices for covariance function and trend function.
Q6. What is the aggregation of the surrogate models?
the cross validation errors of the aggregation is also a quadratic form of the weights (Equation (13)) where C is the same defined in the previous section.
Q7. what is the aggregation of the surrogate models?
(k)|Zn(x) (10)In their formulation, the authors compute the weights of the aggregations by optimizing the PPS of the aggregation under the constraint p ∑i=1 wi = 1.
Q8. What techniques are useful to estimate the errors of a surrogate model?
it is convenient to use re-sampling techniques such as Cross-Validation (CV) (Stone, 1974) and bootstrap (Efron and Tibshirani, 1993) to estimate the predicted errors.