Conceptual Business Process Structuring by Extracting Knowledge from Natural Language Texts

摘要

This article discusses methods of constructing a formalized structure of a subject domain based on analysis of natural language texts, including discovering objects, their properties and related actions, followed by discovering business processes specific to the subject domain and the formation of thesaurus and business processes of the subject domain. At the same time the thesaurus can be changed based on the results of text analysis. Implementation of this approach involves automatic identification of objects mentioned in different documents, determining their properties and relationships, as well as the construction of formalized structure of subject domain by automatic extraction information from natural language texts. The article describes methods of primary text classification in association with their particular subject domain, allowing substantially reduce the number of irrelevant documents for consideration.