Comparison of Group-Contribution and Machine Learning-based Property Prediction Models with Uncertainty Quantification

Adem R. N. Aouichaoui, Resul Al, Jens Abildskov, Gürkan Sin*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Abstract

This study demonstrates the development of three modeling approaches for predicting thermo physical property with the ability to quantify the uncertainty in the prediction. The modeling approaches consist of a classical non-linear group-contribution (GC) model (GCM), Gaussian-Process regression (GPR), and a deep neural network (DNN) all applied to the first-order groups defined by Marrero and Gani as the molecular descriptor. The uncertainty was quantified using different methods: linear error propagation using the parameter covariance matrix for the GCM, the inherent uncertainty quantification of GPR models, and using a probabilistic layer able to learn the distribution of model output sin DNN. The models have been applied to the lower flammability limit (LFL) at 298K. The model performance was evaluated using 5 folds cross-validation to ensure the models were exposed to all data and to detect potential overfitting,—a procedure frequently used with in machine learning. The models obtained produce a good fit to the experimental data when applied to all available data with a coefficient of determination (R2) above 0.9 for all models, a maximum mean absolute error of 0.39 [%-vol], and a maximum mean squared error of 0.51.
Original languageEnglish
Title of host publicationProceedings of the 31th European Symposium on Computer Aided Process Engineering (ESCAPE30)
EditorsMetin Türkay, Rafiqul Gani
Place of PublicationAmsterdam
PublisherElsevier
Publication date2021
Pages755-760
ISBN (Electronic)978-0-323-98325-9
DOIs
Publication statusPublished - 2021
Event31st European Symposium on Computer Aided Process Engineering (ESCAPE 31) - Istanbul, Turkey
Duration: 6 Jun 20219 Jun 2021

Conference

Conference31st European Symposium on Computer Aided Process Engineering (ESCAPE 31)
Country/TerritoryTurkey
CityIstanbul
Period06/06/202109/06/2021
SeriesComputer Aided Chemical Engineering
ISSN1570-7946

Keywords

  • QSPR
  • Deep-Learning
  • Property Prediction
  • Uncertainty Analysis

Fingerprint

Dive into the research topics of 'Comparison of Group-Contribution and Machine Learning-based Property Prediction Models with Uncertainty Quantification'. Together they form a unique fingerprint.

Cite this