Building of knowledge bases of decision support systems using the directed networks of terms during information operations research
| Autor: | Vitaliy V. Tsyganok, Yaroslava Porplenko, Oleh Andriichuk, Dmytro Lande, Oleh Dmytrenko |
|---|---|
| Rok vydání: | 2020 |
| Předmět: |
Marketing
Pharmacology Text corpus Organizational Behavior and Human Resource Management Decision support system Information retrieval Computer science business.industry Strategy and Management Pharmaceutical Science Domain (software engineering) Knowledge base Information space Drug Discovery Information Operations Graph (abstract data type) Relevance (information retrieval) business |
| Zdroj: | Collection "Information Technology and Security". 8:153-163 |
| ISSN: | 2518-1033 2411-1031 |
| DOI: | 10.20535/2411-1031.2020.8.2.222597 |
| Popis: | The relevant issue of information operations research is considered, which includes issues of detection, recognition, analysis, and counteraction. This relevance is due to the significant impact of the information environment on society, social groups and individuals who are in it. The construction of knowledge bases of decision support systems in the information operations research takes place within the method of hierarchical goal dynamic estimating of alternatives. A new approach for the building knowledge bases of decision support systems during information operations research is proposed, which allows reducing the number of appeals to experts through the use of directed weighted networks of terms. Its application saves time and financial resources by reducing expert information usage and also provides an opportunity to identify gaps in knowledge bases. The concept of an information-analytical system for information operations research is developed, which uses directed networks of terms, within which this approach will be implemented. This system allows making recommendations that can be used to analyze and counteract information operations. During automated processing of texts related to the subject domain of the information operation object, there are conducted the formation of text corpora, based on which the construction of networks of terms. Processing of text corpora is performed by computational linguistics, and the construction of a directed network of terms is conducted using an algorithm for constructing a graph of horizontal visibility. It is proposed to determine a sufficient number of texts for representative coverage of the subject domain by stabilizing the ranks of terms when constructing thematic text corpora. An approach for determining the weights of links in the subject domain's directed networks is proposed. The application of the proposed approaches was tested on the example of texts corpora on current topics “Brexit”. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|