D.V. Lande, O.O. Dmytrenko, O.V. Andriichuk,
V.V. Tsyganok, Y.V. Porplenko
Building of directed weighted networks of terms for decision-making support during information operations recognition
// Математичне та ╕м╕тац╕йне моделювання систем. МОДС 2020 : тези допов╕дей П'ятнадцято╖ м╕жнародно╖ науково-практично╖ конференц╕╖ (29 червня . 01 липня 2020 р., м. Черн╕г╕в) / М-во осв╕ти ╕ науки Укра╖ни ; Нац. Акад. наук Укра╖ни ; Академ╕я технолог╕чних наук Укра╖ни ; ╤нженерна академ╕я Укра╖ни та ╕н. - Черн╕г╕в : ЧНТУ, 2020. - P. 147-148.
The study is devoted to automated processing of documents of a subject domain for formation of text corpus, construction of directed weighted networks of terms and using of these networks for building of knowledge bases of decision support systems during information operations recognition.
For forming the text corpus, a tool for professionals in the field of information analysis and information warfare - the content monitoring system "InfoStream" was used. As a result, 135 publications that thematically related to Brexit were downloaded. The computerized processing of text corpus such as the tokenization, lemmatization, stop-words removal, stemming process and terms weighting was made. In this work, for building the undirected networks of terms the Horizontal Visibility Graph algorithm is used.
A new approach for determining the directions and the weights of links in the network of terms that correspond to certain concepts of the considered subject domain is proposed. By its application, an ontological model of the subject domain related to the information operations recognition was built. A new approach for building the knowledge bases of decision support systems during the information operations recognition using the directed weighted networks of terms of subject domain is proposed. Using it for building of knowledge bases allows saving time and financial resources by reducing the use of expert information and make it possible to detect gaps in the knowledge bases of decision support systems.