• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

Book chapter

Information System for Inorganic Substances Physical Properties Prediction Based on Machine Learning Methods

P. 82-85.
Dudarev V.A., Kiselyova N., Stolyarenko A., Dokukin A., Senko O., Ryazanov V., Vashchenko E., Vitushko M., Pereverzev-Orlov V.

ParIS (Parameters of Inorganic Substances) system was developed for predicting inorganic substances physical properties. It is based on the use of machine learning methods to find the relationships between inorganic substances parameters and the properties of chemical elements. The main components of the system are an integrated database system on inorganic substances and materials properties, a subsystem of machine learning and prediction results analysis, a knowledge base and a prediction database. The machine learning subsystem includes programs based on the algorithms developed by the authors of this paper and the algorithms included in the scikit-learn package. The results of the ParIS system application are illustrated by an example of predicting chalcospinels crystal lattice parameter. To get prediction results, only the properties of chemical elements included in the composition of not yet synthesized chalcospinels were used. Moreover, the prediction accuracy was within ± 0.1 Å.