Development of an Ergonomic Assistant Application Based on a Neural Network

Purpose of research is to develop a structured model for training a neural network based on an electronic encyclopedia.

Methods. In this work, methods were used to analyze text articles: 1) TF-IDF is a statistical measure used to assess the importance of terms in a document relative to the corpus of documents. It calculates the relative importance of terms, taking into account the frequency of their appearance in a document or the entire corpus of documents. TFIDF allows you to highlight the key terms that characterize the content of each document. 2) Clustering by k-means++ is a machine learning method used to divide objects into groups (clusters) based on their similarity. This method made it possible to create new categories of articles based on their content. 3) The t-SNE method is a method of visualizing complex multidimensional data in two or three dimensions.

Result. The analysis of text articles using natural language processing methods, such as TF-IDF, and their clustering by the k-means++ method allow you to create new categories of articles based on their content. In this paper, the resulting clustering graph was obtained, which showed well-separated clusters with high accuracy.

Conclusion. TF-IDF values were used to analyze the uniqueness of trigrams in the articles. The trigrams were classified into various uniqueness groups, which allows us to identify areas where articles contain high uniqueness of trigrams. Combinations of different uniqueness groups may indicate different thematic aspects or contexts in the texts. Additionally, using the t-SNE method, a clustering graph was obtained that visually highlights new categories of articles. This graph helps to visualize the structure of clusters and their mutual location in two dimensions. Thus, the study allows a deeper understanding and systematization of the content of articles and highlight the links between them.

1. Ermakov P. D. Issledovanie metodov mashinnogo obucheniya v zadache avtomaticheskogo opredeleniya tonal'nosti tekstov na estestvennom yazyke [Research of machine learning methods in the problem of automatic determination of the tonality of texts in natural language]. Novye informatsionnye tekhnologii v avtomatizirovannykh sistemakh = New Information Technologies in Automated Systems, 2015, no. 18, pp. 600-615.

2. Shokina M. O. Primenenie algoritma k-means++ dlya klasterizatsii posledovatel'- nostei s neizvestnym kolichestvom klasterov [Application of the k-means++ algorithm for clustering sequences with an unknown number of clusters]. Novye informatsionnye tekhnologii v avtomatizirovannykh sistemakh = New Information Technologies in Automated Systems, 2017, no. 2, pp. 160-163.

3. Krizhanovsky A. A. Preobrazovanie struktury slovarnoi stat'i Vikislovarya v tablitsy i otnosheniya relyatsionnoi bazy dannykh [Transformation of the structure of a Wiktionary dictionary entry into tables and relational database relationships], Available at: http://scipeople.com/publication/1000231.

4. Chastikova V.A., Ostapov D.S. Primenenie metodov klasterizatsii dlya povysheniya tochnosti raboty neironnykh sete [Application of clustering methods to improve the accuracy of neural networks]. Sovremennye problemy nauki i obrazovaniya = Modern Problems of Science and Education, 2015, no. 1-1.

5. Soboleva E. D., Popova I. A., Popova A. A. Vizualizatsiya mnogomernykh naborov dannykh pri pomoshchi algoritmov snizheniya prostranstva priznakov pca i t-sne [Visualization of multidimensional data sets using algorithms for reducing the feature space of pca and t-sne]. StudNet Publ., 2020, vol. 3, no. 11, pp. 982-1004.

6. Lozbinev F.Yu., Sazonova A.S., Tishchenko A.A., Leonov Yu.A. Prognozirovanie zhivuchesti mul'tiservisnoi korporativnoi seti svyazi [Forecasting the survivability of a multiservice corporate communication network]. Vestnik Bryanskogo gosudarstvennogo tekhnicheskogo universiteta = Bulletin of the Bryansk State Technical University, 2017, no. 4 (57), pp. 144-150.

7. Kazakov Yu.M., Tishchenko A.A., Kuzmenko A.A., Leonov Yu.A., Leonov E.A. Metodologiya i tekhnologiya proektirovaniya informatsionnykh sistem [Methodology and technology of information systems design]. Moscow, FLINT Publ., 2018, 136 p.

8. Lagerev A.V., Sazonova A.S., Filippov R.A. Model' otsenki sotsial'nodemograficheskogo potentsiala i ego vliyanie na strukturu vysshego professional'nogo i poslevuzovskogo obrazovaniya v regione [A model for assessing socio-demographic potential and its impact on the structure of higher professional and postgraduate education in the region]. Informatsionnye sistemy i tekhnologii = Information Systems and Technologies, 2012, no. 3 (71), pp. 72-77.

9. Leonov Yu.A., Leonov E.A., Zueva A.S., Sazonova A.S. Poisk optimal'nykh tekhnologicheskikh protsessov s ispol'zovaniem algoritmov evristicheskogo poiska [Search for optimal technological processes using heuristic search algorithms]. Vestnik Bryanskogo gosudarstvennogo tekhnicheskogo universiteta = Bulletin of the Bryansk State Technical University, 2017, no. 4 (57), pp. 122-127.

10. Tanygin M. O., Ahmad A.A., Kazakova O. V., Golubov D. A. Recursive Algorithm for Forming Structured Sets of Information Blocks that Increase the Speed of Performing Procedures for Determining Their Source. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta = Proceedings of the Southwest State University. 2023; 27(1): 73-91 (In Russ.). https://doi.org/10.21869/2223-1560-2023-27-1-73-91.

11. Averchenkova E.E., Averchenkov A.V., Leonov YU.A., Kravcov D.V., Filippova L.B., Leonov E.A. Ekonomicheskaya bezopasnost' v regional'nyh social'no-ekonomicheskih sistemah [Economic security in regional socio-economic systems]. Moscow, FLINTA Publ., 2019, 157 p.

12. Filippov R.A., Filippov R.A., Filippova L.B., Averchenkov A.V., Sazonova A.S., SHeptunov S.A. Razrabotka matematicheskoj modeli informacionnoj sistemy dlya inventarizacii i monitoringa programmnogo i apparatnogo obespecheniya na osnove metodov nechetkoj logiki [Development of a mathematical model of an information system for inventorying and monitoring software and hardware based on fuzzy logic methods]. Kachestvo. Innovacii. Obrazovanie = Quality. Innovation. Education, 2018, no. 7 (158), pp. 105-112.

13. Averchenkov V.I., Spasennikov V.V., Filippov R.A. Issledovanie tochnosti pozicionirovaniya ob"ektov pri opticheskoj mikroskopii s upravleniem cherez Internet [Investigation of the accuracy of object positioning in optical microscopy with control via an Internet]. Vestnik Bryanskogo gosudarstvennogo tekhnicheskogo universiteta = Bulletin of the Bryansk State Technical University, 2012, no. 1 (33), pp. 125-130.

14. Sazonova A.S., Filippova L.B., Filippov R.A. Ocenka innovacionnogo potenciala regiona [Assessment of the innovative potential of the region]. Vestnik Voronezhskogo gosudarstvennogo universiteta inzhenernyh tekhnologij = Bulletin of the Voronezh State University of Engineering Technologies, 2017, vol. 79, no. 2 (72), pp. 273-279.

15. ZHadaev D.S., Kuz'menko A.A., Spasennikov V.V. Osobennosti nejrosetevogo analiza urovnya podgotovki studentov v processe adaptivnogo testirovaniya ih professional'nyh kompetencij [Features of neural network analysis of the level of training of students in the process of adaptive testing of their professional competencies]. Vestnik Bryanskogo gosudarstvennogo tekhnicheskogo universiteta = Bulletin of the Bryansk State Technical University, 2019, no.2 (75), pp. 90-98.

16. Kondratenko S.V., Kuz'menko A.A., Spasennikov V.V. Metodologiya ocenki deyatel'nosti operatorov v cheloveko-mashinnyh sistemah [Methodology for evaluating operator activity in human-machine systems]. Vestnik Bryanskogo gosudarstvennogo tekhnicheskogo universiteta = Bulletin of the Bryansk State Technical University, 2017, no. 1 (54), pp. 261-270.

17. KMeansTrainer Klass. Available at: https://docs.microsoft.com/ru-ru/dotnet/api/microsoft.ml.trainers.kmeanstrainer?view=ml-dotnet, svobodnyj.

18. Karamysheva N. S., Svishchev D. S., Popov K. V., Zinkin S. A. Implementation of Agent-Based Metacomputersystems and Applications. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta = Proceedings of the Southwest State University. 2022; 26(1): 148-171 (In Russ.). https://doi.org/10.21869/2223-1560-2022-26-1-148-171

19. Milostnaya N. A. Stability Study of a Neuro-Fuzzy Output System Based on Ratio Area Method. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta = Proceedings of the Southwest State University. 2021; 25(3): 70-85 (In Russ.). https://doi.org/10.21869/2223-1560-2021-25-3-70-85.

20. Martyshkin А. I., Kiryutkin I. А., Merenyasheva Е. А. Autotesting an Embedded Reconfigurable Computing System. Izvestiya Yugo-Zapadnogo gosudarstvennogo universiteta = Proceedings of the Southwest State University. 2023; 27(1): 140-152 (In Russ.). https://doi.org/10.21869/2223-1560-2023-27-1-140-152.