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

DOI10.1016/J.CPC.2021.108220zbMATH Open1528.65012arXiv2012.02704OpenAlexW3212467745MaRDI QIDQ6156985FDOQ6156985

Owen Ren, Manabu Ihara, Sergei Manzhos, Dmitry Voytsekhovsky, Mohamed Ali Boussaidi

Publication date: 19 June 2023

Published in: Computer Physics Communications (Search for Journal in Brave)

Abstract: We present a Python implementation for RS-HDMR-GPR (Random Sampling High Dimensional Model Representation Gaussian Process Regression). The method builds representations of multivariate functions with lower-dimensional terms, either as an expansion over orders of coupling or using terms of only a given dimensionality. This facilitates, in particular, recovering functional dependence from sparse data. The code also allows for imputation of missing values of the variables and for a significant pruning of the useful number of HDMR terms. The code can also be used for estimating relative importance of different combinations of input variables, thereby adding an element of insight to a general machine learning method. The capabilities of this regression tool are demonstrated on test cases involving synthetic analytic functions, the potential energy surface of the water molecule, kinetic energy densities of materials (crystalline magnesium, aluminum, and silicon), and financial market data.

Full work available at URL: https://arxiv.org/abs/2012.02704

zbMATH Keywords

machine learningGaussian process regressionhigh dimensional model representationmultivariate functiondata imputation

Mathematics Subject Classification ID

Numerical smoothing, curve fitting (65D10) Multidimensional problems (41A63) Statistical sampling theory and related topics (62D99)

Cites Work

• Title not available (Why is that?)
• Title not available (Why is that?)
• General foundations of high-dimensional model representations
• Efficient implementation of high dimensional model representations

Cited In (1)

• High dimensional model representations generated from low dimensional data samples. I: mp-cut-HDMR

This page was built for publication: 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

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6156985)