AUTOMATIC DOCUMENTATION AND MATHEMATICAL LINGUISTICS最新文献

Some possible mathematical models and methods of intelligent data analysis (IDA) for the field of evidence-based medicine (EBM) are discussed. Two critically significant limitations for the application of traditional (statistics-based) EBM approach are considered: work with open subject areas and with small (statistically nonsignificant) collections of analyzed data. A special class of IDA methods based on the computer-oriented formalization of causal similarity heuristics using logical and algebraic means is presented. Options for clarifying the concept of evidence-based are proposed, which allow the indicated limitations of the traditional EBM approach to be circumvented. Some practically significant characteristics of this variant of the use of artificial intelligence methods in the tasks of evidence-based medicine are discussed.

The concepts of a scientific library and a network scientific library are clarified. The role of scientific network libraries in solving issues of information support for scientific research and specific scientists is shown. It is noted that the Library of Natural Sciences of the Russian Academy of Sciences (BEN RAS) is a classic online scientific library, with more than 50 divisions. In this paper, the network technology of an automated library and information system (ALIS) is understood as a single automated library complex, which includes a geographically ungrouped system of terminals combined into a single system by means of connections, communication equipment, software, data transmission protocols, and control and computing tasks. The main requirements for network technologies include their ease of use; the ability to access other networks and systems and the internet; high data transfer speed; and high information security. The generations of the development of the ALIS are considered, and it is noted that the attention is attracted by open source ALIS. The basic principles of building an ALIS network library are formulated, which are presented in the form of three blocks: methodological, technological, and linguistic. A general model of network ALIS has been developed using the example of BEN RAS, which includes the main technological blocks: management of all processes; acquisition of information funds; processing of all information resources entering the AIS and resources of its own generation; storage of information resources: traditional and electronic, to be stored and used in information service processes; and user services. The key connecting element is the information need both for the formation of information resources included in the ALIS and for the organization of an information support system for scientific research. The peculiarities of the network structure are that the information support unit uses different classification systems and different levels of categorization to match user requests more accurately, whereas the formation of information arrays takes place, as a rule, at the upper level of one or another classification system chosen as the main one.

The paper introduces a methodology for extracting opinion aspects from textual content by identifying the customer-evaluated parameters regarding a given object. These parameters form the foundation for shaping the customer’s attitudes toward the product or service. The proposed approach leverages topic modeling tools to delineate classes of vocabulary exhibiting semantics aligned with the parameters influencing the customer’s opinion about the object. Our study specifically explores the application of the BERTopic model as a topic modeling tool to address this challenge. The outlined methodology encompasses several sequential steps, including the preprocessing of textual data involving the removal of stopwords, conversion to lowercase characters, and lemmatization. Additionally, special consideration is given to the distinct lexical manifestations of opinion aspects, obtained as a result of the extraction of nominal, verbal, and adjectival single- and multicomponent phrases from the corpus. Subsequently, the corpus sentences are represented as vectors in a feature space expressed by the extracted words and phrases. The final step involves the application of topic modeling using the BERTopic model on the customer review corpus, utilizing the vector representations of corpus sentences. The experimental inquiry is conducted on a domain-specific Russian-language corpus comprising customer feedback on airline services gathered from customer review websites. The resultant topic distribution is then juxtaposed against a manually constructed conceptual model of the domain. The comparative analysis reveals that the automatic topic distribution aligns with the conceptual structure of the domain, demonstrating a precision of 0.955 and a recall of 0.875. These findings affirm the efficacy of employing the BERTopic model to address the problem of the corpus-based mining of opinion aspects.

The article deals with extensions of empirical regularities (ER) for r periods, where r > 1. Various types of ER constitute a lattice. Sequences of ER types correspond in a one-to-one manner to sequences of empirical modalities which form partially ordered semigroups. Syntactic and semantic types of ER are defined for the elements of these semigroups. The equality of semantic ER types that represent two different methods of extending fact bases determine the pragmatic condition for the acceptance of ER hypotheses.

Articles scattered in the VINITI abstract database dated 2020–2022 are studied using the example of three thematic fragments: chemistry and chemical technologies, mechanical engineering, and metallurgy and welding. The mathematical formulas of Bradford and Vickery were used to calculate the number of journals in the third scattering zone. It is argued that the Bradford–Vickery law of publication scattering allows us to predict the minimum number of journals that cover more than 90% of articles quantity on a certain topic. The optimization of journal selection is reduced to the solution of the well-known knapsack problem. It is proposed to rank journals by thematic relevance. Unlike the Bradford model, intended to select the most productive journals, the proposed model is designed to select the most relevant journals.

Hierarchical, descriptive, and faceted methods of constructing knowledge classifiers and also the classification of knowledge using folksnomies are described. Mathematical models to formalize the listed methods of constructing classifiers are presented. A general description of the classifiers of the Russian Science Citation Index, Code of State Categories Scientific and Technical Information, and Universal Decimal Classification is given. The Chinese experience of classifying scientific publications and the American experience of classifying patent documents are analyzed. An alternative possibility for classifying scientific publications using clustering with bibliometric indicators and using keywords is indicated. A review of the main methods and means of comparing knowledge classifiers using qualifiers and expert competencies is carried out. A new approach to the compilation of knowledge classifiers based on their reduction to oriented trees and the construction of homomorphisms between these graphs is proposed.

This work is devoted to the development of a modification of the particle collision algorithm (PCA), which provides an approximate solution to the traveling salesman problem. The resulting modification was tested on a number of well-known tasks and demonstrated greater accuracy and efficiency than its analogues.

Trends in the implementation of digital twins (DTs) of real-world objects for the effective and safe management of complex sociotechnical systems and processes have been identified. Their composition and their share of the IT market are discussed. The features of the formation of data centers on hardware and software platforms from the most well-known companies in the IT industry are considered. It is proposed that DT technologies be used in creating conditions for a high-quality and mobile educational process that takes into account the individual trajectories of personal growth of students. The concept of a digital student profile (DSP) is presented, which is convenient for obtaining a quick quantitative assessment of personal skills and abilities, which is convenient when entering the labor market. In an educational environment, DSPs can be used to analyze the academic achievements, behavior and needs of students. Criteria for the formation of DSPs as a multidimensional quantitative assessment are proposed. Using the example of the course Information Technologies in Science and Education, a criteria-based convolutional assessment of the mastery of the material covered is constructed, characterizing the achievements of a particular student, allowing the use of methods and means of data mining. The creation and application of DTs and DPSs is an important step in the development of a new paradigm of education, with the help of which it is possible to significantly improve the quality of learning and provide participants in the educational process with individual support and assistance in achieving their unique life goals.

This article deals with the potential of digitalization to improve participants’ skills in the agri-food market. The topic of improving technical mechanisms and software technologies aimed at the qualitative development of e-education in Russia is being updated. The problems of providing the agri-food market with skilled personnel are analyzed. The problems of terminological uncertainty concerning the concepts of digitalization and distance education in computer sciences (informatics), information sciences, and other related scientific fields that have a negative impact on the overall wide turnover of scientific research, are touched upon. The authors propose a single generalized nonhierarchical technological model of the organization of e-education, schematically describing a set of software technologies in the form of a cloud telecommunications system. Particularly, it is recommended to use multicasting technologies and adaptive network means of signal delivery to promote educational content in the form of internet television. The basic integrative network tools and protocols for the specific organization of electronic educational activities via internet television are technically annotated. The conclusions briefly and conceptually reflect the course of the research, as well as the logic of the material analyzed by the authors, summarizing the research work done and to provide some recommendations for the operation of multicasting technology in the form of internet television in Russia.