I describe a machine learning approach to train on and predict thermal rate constants over a large temperature range. Cluster analysis is done first and Gaussian process regression (GPR) is trained on each cluster. Each cluster is defined by the magnitude of the correction to the Eckart transmission coefficient needed to bring the rate constant into agreement with accurate calculated ones. Owing to small number of such calculations, training is done on the full dataset for each cluster. Testing is done by inputing hundreds of random values of the descriptors (within reasonable bounds). The new training strategy is applied to predict the rate constants of the O(3P)+HCl reaction on the 3A′and 3A′′potential energy surfaces. This reaction was recently presented as a “stress test” for the ring polymer molecular dynamics method. The database and python software can be downloaded at https://scholarblogs.emory.edu/bowman/gpr-for-rate-constants/

A machine learning approach for rate constants: Clustering, training, and predictions for the O(3P)+HCl --> OH+Cl reaction

ACS Spring Meeting 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.