Comments (3)
With only 3 inputs, I would say you have too many training point and this can lead to ill-conditionned kriging matrix (kriging does not like too many points too close from each other). I would suggest to try with less training points. As a rule of thumb for a small input dimension, 10*dimension training points covering your input space should work fine.
from smt.
If you want to keep your 815 points in your training DOE, the other idea is to modify the correlation matrix by adding 'a small term' on the diagonal to improve the conditioning of the matrix (see this paper : https://www.asc.ohio-state.edu/statistics/comp_exp/jour.club/AndrianakisChallenor_CSDA2012.pdf).
The value of the nugget is chosen in file krg_based.py line 170; so if you modify it to :
nugget = 100.0 * MACHINE_EPSILON
instead of
nugget = 10.0 * MACHINE_EPSILON
You won't have anymore the error.
from smt.
@relf Thank you for the comment.
As you suggested, I reduced the number of points to 30 and the training was successfull. I will keep this point in mind.
@NatOnera Thank for the suggestion and the paper link. I will try this way.
from smt.
Related Issues (20)
- Does SMT support parallel computing or GPU computing? HOT 1
- SMT version 1.3 or 2.0? HOT 2
- Can't Run Example EGO Code HOT 2
- Kriging surrogate model in case of multiple outputs HOT 5
- Sampling Method sometimes gets stuck HOT 2
- pyDOE2 non-functional with python 3.12, appears to be unsupported. HOT 5
- Question on drawing drawing sample from MFK GP posterior. HOT 3
- Any interest in putting this on conda-forge? HOT 3
- Zero predicted variance while prediction differs from ground truth HOT 1
- Adaptive sampling methods for Kriging models and its variants including multi-fidelity Kriging HOT 1
- AttributeError: 'DesignSpace' object has no attribute '_cs' HOT 4
- InactiveHyperparameterSetError encountered when solving the HierarchicalGoldstein problem. HOT 3
- data-driven Multi-Fidelity Kriging HOT 1
- ValueError by passing an int instead of a float HOT 3
- SMT 2.5 is not available on conda-forge HOT 2
- Error from SMT --> ValueError: setting an array element with a sequence. HOT 4
- Question on new syntax to train GENN (JENN) models HOT 6
- Meta: add "Discussions" to GitHub repo? HOT 1
- Saving models without using pickle? HOT 1
- Number of parameters can be handled by KPLSK HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from smt.