О.В. Золотарёв

In this paper we describe a semantic approach to visualization of 3D cyberspace of scientific papers and their research front using web-based 3D graphic. The most cited and signifi cant documents in this cyberspace are represented by spheres of a large size, and the distance between documents is proportional to their semantic similarity. A new measure of semantic similarity of documents is proposed that is determined by the maximum correlation between explicit and implicit connectivity of the documents. A new science contextual citation index (SCCI) that is defined by a correlation maximum with a science citation index (SCI) is proposed and implemented. SCCI can more accurately measure scientific impact, find significant documents and evaluate new articles with zero SCI. The significant similar articles confi rm each other and form clusters in the cyberspace. The research front exists as a set of such clusters. The proposed cyberspace implemented by WebVR and interactive 3D graphics can be considered as a dynamic learning environment that is convenient for discovering new significant articles, ideas and trends.

The paper discusses the problems of semantic modeling techniques for automated detection of hierarchical, synonymous and associative relationships from online texts and the construction of linguistic and statistical portraits of various subject areas. The study is based on the hypothesis that the more general terms have more associative relations. The involvement of associative relationships for the definition of the full meaning is revealed by the context of the environments that gives you the ability to automate the process of differentiating between values and knowledge extraction from texts. The solution is based on an integrated approach that combines statistical methods, corpus linguistics and distributional semantics, and is implemented in a technology which involves the development of linguo-statistical mechanisms for the formation of associative-hierarchic portrait of the subject area (AHPSA), which is a dictionary of important terms of the subject area, elements of which are connected by the associative and hierarchical relationships. Work is carried out on the basis of the analysis of different subject areas, in particular, Autonomous Unmanned Underwater Vehicle (AUUV).

This article discusses the approaches for developing onthologies of the subject area based on the analysis of natural language texts. The article defines the concepts of ontology, semantic search, and describes the structure of the ontology, different aspects of the formation of a search query, the features of information extraction from natural language texts.

This article discusses the methods for identifying duplicate documents to prevent their inclusion into collections; it also analyses the approaches for fi nding near-duplicate documents based on the method of shingles to determine spam e-mail, search of plagiarism, to clean collections’ documents from duplicates. This paper also presents analysis methods and parameter selection of shingle algorithm, criteria of the checksums (signatures) selection. There was developed a program for identifying duplicates, and proposed the criteria for selecting the optimization algorithm of shingles with using MinHash and supershingles algorithm.

The method for detection of intertextual relations by finding implicit links using linguistic and statistical methods is suggested. Intertextuality is a presence in one text of elements and ideas from other texts. The possibility of identifying of crosslanguage migration of terms and ideas for prognosis and determination of ideological trajectories is demonstrated. A new text similarity measure is suggested. The measure was tested using collection of scientific documents. The measure was improved by maximizing correlation between explicit and implicit links. A method for documents clustering according to the measure of text similarity is suggested. The possible application of the proposed measure for analysis of extremist texts from the Internet is suggested.

The paper deals with the analysis of the degree of terrorist threats on the Internet based on the analysis of natural language texts. The article describes the tasks of identifying terrorist groups in the Internet environment. It also analyzes natural language texts on the Internet. For the analysis of texts we use the technique of extended semantic networks. The paper presents examples of the results of the semantic analysis of texts.

An important current trend in the assessment of the work of scientists is not only to account the formal references in published articles, but also in the analysis of the content of scientifi c texts. The proposed approach to the analysis of texts is to identify the informal (implicit) references in text to the author’s ideas. The developed method of detecting the implicit links provides a high correlation of implicit links with the formal links. In the course of further studies there is planned to develop measures to improve the correlation of the explicit and implicit links. This will allow you to more accurately and objectively assess not only the works of scientists, but also to the interaction between ideas in different documents, even in those that rarely use the formal links, such as, for example, in the online Internet texts.