The invention discloses a
machine learning-based
antioxidant prediction system for a
resveratrol dimer, and belongs to the technical field of biological
antioxidant experiments.The
machine learning-based
antioxidant prediction system comprises the steps of determining dominant conformation and a
reaction mechanism, summarizing and obtaining an antioxidant structure-function relationship and an influence rule of the
resveratrol dimer, and predicting the antioxidant structure-function relationship of the
resveratrol dimer. The anti-oxidation mechanism module determines an energy channel and an anti-oxidation mechanism of the resveratrol dimer in the process of removing different free radicals, the reliability
verification module performs an experiment based on theoretical
simulation and verifies the reliability of calculation of the anti-oxidation ability of the resveratrol dimer, and the prediction guidance module determines the anti-oxidation ability of the resveratrol dimer based on the theoretical
simulation and the experimental
verification result. According to the method, a resveratrol
biopolymer antioxidant
database is constructed, different
machine learning models are adopted for screening, an excellent antioxidant is predicted and guided to be designed according to a
screening result, and the method has important practical application value for deep development and utilization of three types of resveratrol biopolymers.