Academic Bibliography

 Search 200 years of publications by Ghent University researchers.
Advanced search
2 files | 7.24 MB
Add to list
  1. Search results
  2. Exploring machine learning methods fo...

Exploring machine learning methods for absolute configuration determination with vibrational circular dichroism

Tom Vermeyen (UGent) , Jure Brence, Robin Van Echelpoel, Roy Aerts, Guillaume Acke (UGent) , Patrick Bultinck (UGent) and Wouter Herrebout
(2021) PHYSICAL CHEMISTRY CHEMICAL PHYSICS. 23(35). p.19781-19789
Author
Organization
Project
Abstract
The added value of supervised Machine Learning (ML) methods to determine the Absolute Configuration (AC) of compounds from their Vibrational Circular Dichroism (VCD) spectra was explored. Among all ML methods considered, Random Forest (RF) and Feedforward Neural Network (FNN) yield the best performance for identification of the AC. At its best, FNN allows near-perfect AC determination, with accuracy of prediction up to 0.995, while RF combines good predictive accuracy (up to 0.940) with the ability to identify the spectral areas important for the identification of the AC. No loss in performance of either model is observed as long as the spectral sampling interval used does not exceed the spectral bandwidth. Increasing the sampling interval proves to be the best method to lower the dimensionality of the input data, thereby decreasing the computational cost associated with the training of the models.
Keywords
Physical and Theoretical Chemistry, General Physics and Astronomy, INFRARED-SPECTRA, CLASSIFICATION, IDENTIFICATION, SPECTROSCOPY, SIMULATION, PREDICTION, REGRESSION, SYSTEMS, SHIFTS, RAMAN

Downloads

  • Download
    main article.pdf
    • full text (Accepted manuscript)
    • |
    • open access
    • |
    • PDF
    • |
    • 3.00 MB
  • Download
    (...).pdf
    • full text (Published version)
    • |
    • UGent only
    • |
    • PDF
    • |
    • 4.24 MB

Citation

Please use this url to cite or link to this publication:

MLA
Vermeyen, Tom, et al. “Exploring Machine Learning Methods for Absolute Configuration Determination with Vibrational Circular Dichroism.” PHYSICAL CHEMISTRY CHEMICAL PHYSICS, vol. 23, no. 35, 2021, pp. 19781–89, doi:10.1039/d1cp02428k.
APA
Vermeyen, T., Brence, J., Van Echelpoel, R., Aerts, R., Acke, G., Bultinck, P., & Herrebout, W. (2021). Exploring machine learning methods for absolute configuration determination with vibrational circular dichroism. PHYSICAL CHEMISTRY CHEMICAL PHYSICS, 23(35), 19781–19789. https://doi.org/10.1039/d1cp02428k
Chicago author-date
Vermeyen, Tom, Jure Brence, Robin Van Echelpoel, Roy Aerts, Guillaume Acke, Patrick Bultinck, and Wouter Herrebout. 2021. “Exploring Machine Learning Methods for Absolute Configuration Determination with Vibrational Circular Dichroism.” PHYSICAL CHEMISTRY CHEMICAL PHYSICS 23 (35): 19781–89. https://doi.org/10.1039/d1cp02428k.
Chicago author-date (all authors)
Vermeyen, Tom, Jure Brence, Robin Van Echelpoel, Roy Aerts, Guillaume Acke, Patrick Bultinck, and Wouter Herrebout. 2021. “Exploring Machine Learning Methods for Absolute Configuration Determination with Vibrational Circular Dichroism.” PHYSICAL CHEMISTRY CHEMICAL PHYSICS 23 (35): 19781–19789. doi:10.1039/d1cp02428k.
Vancouver
1.
Vermeyen T, Brence J, Van Echelpoel R, Aerts R, Acke G, Bultinck P, et al. Exploring machine learning methods for absolute configuration determination with vibrational circular dichroism. PHYSICAL CHEMISTRY CHEMICAL PHYSICS. 2021;23(35):19781–9.
IEEE
[1]
T. Vermeyen et al., “Exploring machine learning methods for absolute configuration determination with vibrational circular dichroism,” PHYSICAL CHEMISTRY CHEMICAL PHYSICS, vol. 23, no. 35, pp. 19781–19789, 2021.
@article{8722174,
  abstract     = {{The added value of supervised Machine Learning (ML) methods to determine the Absolute Configuration (AC) of compounds from their Vibrational Circular Dichroism (VCD) spectra was explored. Among all ML methods considered, Random Forest (RF) and Feedforward Neural Network (FNN) yield the best performance for identification of the AC. At its best, FNN allows near-perfect AC determination, with accuracy of prediction up to 0.995, while RF combines good predictive accuracy (up to 0.940) with the ability to identify the spectral areas important for the identification of the AC. No loss in performance of either model is observed as long as the spectral sampling interval used does not exceed the spectral bandwidth. Increasing the sampling interval proves to be the best method to lower the dimensionality of the input data, thereby decreasing the computational cost associated with the training of the models.}},
  author       = {{Vermeyen, Tom and Brence, Jure and Van Echelpoel, Robin and Aerts, Roy and Acke, Guillaume and Bultinck, Patrick and Herrebout, Wouter}},
  issn         = {{1463-9076}},
  journal      = {{PHYSICAL CHEMISTRY CHEMICAL PHYSICS}},
  keywords     = {{Physical and Theoretical Chemistry,General Physics and Astronomy,INFRARED-SPECTRA,CLASSIFICATION,IDENTIFICATION,SPECTROSCOPY,SIMULATION,PREDICTION,REGRESSION,SYSTEMS,SHIFTS,RAMAN}},
  language     = {{eng}},
  number       = {{35}},
  pages        = {{19781--19789}},
  title        = {{Exploring machine learning methods for absolute configuration determination with vibrational circular dichroism}},
  url          = {{http://doi.org/10.1039/d1cp02428k}},
  volume       = {{23}},
  year         = {{2021}},
}
Altmetric
View in Altmetric
Web of Science
Times cited: