Jump to main content
Jump to site search

Issue 82, 2020
Previous Article Next Article

Hybridizing physical and data-driven prediction methods for physicochemical properties

Author affiliations

Abstract

We present a generic way to hybridize physical and data-driven methods for predicting physicochemical properties. The approach ‘distills’ the physical method's predictions into a prior model and combines it with sparse experimental data using Bayesian inference. We apply the new approach to predict activity coefficients at infinite dilution and obtain significant improvements compared to the physical and data-driven baselines and established ensemble methods from the machine learning literature.

Graphical abstract: Hybridizing physical and data-driven prediction methods for physicochemical properties

Back to tab navigation

Supplementary files

Article information


Submitted
01 Aug 2020
Accepted
10 Sep 2020
First published
10 Sep 2020

Chem. Commun., 2020,56, 12407-12410
Article type
Communication

Hybridizing physical and data-driven prediction methods for physicochemical properties

F. Jirasek, R. Bamler and S. Mandt, Chem. Commun., 2020, 56, 12407
DOI: 10.1039/D0CC05258B

Social activity

Search articles by author

Spotlight

Advertisements