Extraction of experimental data on chemical kinetics from open sources in the Internet

Abstract

In article process of intellectual search and extraction of experimental data on chemical kinetics from open sources in the Internet is considered. Approach to the organization of the above-stated process which includes the sequence of the following stages is offered: forming of a computer corpus of documents electronic versions of scientific journals, as in open access, and commercial journals, development of a domain ontology for creation of the thesaurus, intellectual search and forming of the electronic corpus of bibliographic links, the intellectual analysis of bibliographic links on the basis of cluster analysis, downloading, conversion to a test format and document classification by a Bayesian neural network, extraction of data and their warehousing. The technologies developed by authors allow to work with English-language and Russian-language texts. The algorithm automatic classifications uses three methods of a clustering. Automatic classification of publications uses specially constructed three-level information basis which breaks sets of articles on the nature of contents. On the basis of the received splitting the case of electronic documents which contains experimental data on reaction activity of organic compounds in chemical reactions is under construction. The retrieved characteristics of chemical reactions remain in the developed database for further use.