Random Sampling High Dimensional Model Representation Gaussian Process Regression (RS-HDMR-GPR) for representing multidimensional functions with machine-learned lower-dimensional terms allowing insight with a general method.
Owen RenMohamed Ali BoussaidiDmitry A. VoytsekhovskyManabu IharaSergei ManzhosPublished in: Comput. Phys. Commun. (2022)
Keyphrases
- gaussian process regression
- random sampling
- high dimensional
- sampling algorithm
- parameter space
- lower dimensional
- prior information
- em algorithm
- similarity measure
- pattern recognition
- prior knowledge
- low dimensional
- closed form
- energy function
- dimensionality reduction
- probabilistic model
- active learning
- data sets
- input data
- model selection
- supervised learning
- sample size
- probability distribution
- dimension reduction
- k means
- data analysis
- image processing
- machine learning