Machine Learning Prediction of the Internal Conversion Rate Constant
Year: 2025
Type: article
Abstract: For the first time, a model for predicting the internal conversion rate constant (S1→S0) was developed using the machine learning method «random forest», based on a series of porphyrins with various substituents. The model was trained on 1058 molecules, while 256 molecular structures were utilized for testing. The constructed model employs molecular geometric parameters as input data, including bond lengths, as well as valence and dihedral angles. The results obtained indicate a correlation between the geometric parameters of the molecule and the internal conversion rate constant. The primary molecular descriptors for the studied porphyrins are the bond lengths and angles within the porphyrin macrocycle.
Language: English
Subfield: Materials Chemistry
Field: Materials Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7128219084
