Knowledge extraction from the problem-oriented data warehouses on chemical kinetics and thermochemistry

This paper discusses the technology of extracting the chemical knowledge from the structured electronic sources, problem-oriented systems of science data analytics, and methods of science data analysis. Application of the feed-forward artificial neural network for predicting the reactivity of a compound (the classical potential barrier) in reactions of hydrocarbons with hydrogen atom in solution is presented. Empirical indexes of reactionary centers for groups of such reactions had identified. The artificial neural network is learned using a set of the experimental thermochemical and kinetic data. The artificial neural network has predicted classic potential barrier for hydrogen atom in reactions with hydrocarbons in solution at a temperature of 298 K with satisfactory accuracy. Special attention is placed the use of fuzzy knowledge base to predict the classical potential barrier and to calculate the rate constants of the bimolecular radical reactions of phenyl radical with hydrocarbons in the liquid phase on the experimental data. The hybrid algorithm of calculation of rate constants of bimolecular radical reactions on the experimental thermochemical data and an empirical index of the reactionary center is offered. This algorithm uses a fuzzy knowledge base or an artificial neural network for prediction of classical potential barrier of bimolecular radical reactions at some temperature, a database of experimental characteristics of reaction and Arrhenius's formula for calculation of rate constant. Results of prediction of the classical potential barrier are discussed. Key-Words: chemical kinetic, thermochemistry, data warehouse, data science analytic, expert system, artificial neural network, fuzzy logic. radical reaction, activation energy, classical potential barrier, reactionary center.