Predicting the Probability of Organic Chemical Reactions Occurring

The thesis is devoted to the development of a web application for predicting the probability of organic chemical reactions. The main task of the work is to create an easy-to-use and effective service that will allow synthetic chemists to validate chemical reactions for the probability of their passage. In this paper, an overview of existing methods for predicting the success of reactions is carried out, algorithms for calculating signs for chemicals are considered, including modern neural network approaches, methods based on the calculation of physico-chemical data and vector representation of molecules using empirically identified patterns. In addition, a new approach has been proposed to train a model for predicting the probability of chemical reactions. The result of the work is a web application that allows you to train a probability prediction model for a user-defined reaction. Upon request, the service returns the probability of a chemical reaction, which was transmitted by the user. Research in this direction will allow scientists to increase the proportion of successful experiments and provide an opportunity to discover new ways of synthesizing organic compounds that were not previously known.