Received 02.09.2024, Revised 22.11.2024, Accepted 26.12.2024

Method of multi-purpose term search in the terminology database

Andriy Yarovyi, Dmytro Kudriavtsev

This study investigated the method of multi-purpose term search in a terminological knowledge base, which is based on semantic analysis and the use of modern natural language processing methods. The study considered the key factors affecting the search efficiency, including the structure of data organisation, data format and parameters, and sample size. Particular focus was placed on the semantic similarity between terms, which allows increasing the search accuracy by using vector representations and the Louvain algorithm. The study also described the use of cosine similarity to quantify the similarity between terms. Furthermore, the search process was optimised by filtering relevant databases and dynamically identifying relevant terms using the modularity metric. A comparative analysis of existing methods for searching for terms by the identified factors was conducted. The study noted the advantages and disadvantages of using the Louvain algorithm in comparison with the search algorithms in graph data structures. A series of experiments were conducted on data samples, including dictionary, graph, and network data structures. The study analysed the use of logistic constraints for searching in network data structures and noted the possibility of optimisation due to uniform and dynamic data distribution. Experimental results showed the effectiveness of using a combination of the Louvain algorithm and network data structures in terminological knowledge bases. Examples of the scope of application of this method in information technologies for searching and processing text data were given. A software architecture scheme with the use of a software interface and the possibility of integration for web applications in the form of a package or library was developed. The proposed approach demonstrates effectiveness in the context of intelligent decision support systems and automated chatbots, which makes it particularly useful for industries where access to accurate professional terms is critical. A basic version of the software interface for using this method in information technologies for searching and analysing data for use in search engines was developed

terminological knowledge base; semantic similarity; Louvain algorithm; vector representations; natural language processing
20-28
Yarovyi, A., & Kudriavtsev, D. (2024). Method of multi-purpose term search in the terminology database. Information Technologies and Computer Engineering, 21(3), 20-28. https://doi.org/10.63341/vitce/3.2024.20

References

[1] Abdykerimova, L., Abdikerimova, G.B., Konyrkhanova, A., Nurova, G., Bazarova, M., Bersugir, M., Kaldarova, M., & Yerzhanova, A. (2024). Analysis of the emotional coloring of text using machine and deep learning methods. International Journal of Electrical and Computer Engineering (IJECE), 14, article number 3055. doi: 10.11591/ijece. v14i3.pp3055-3063.

[2] Baqal, H., & Sidiq, M. (2024). Graph databases: Revolutionizing database design and data analysis. Current Journal of Applied Science and Technology, 43, 45-56. doi: 10.9734/cjast/2024/v43i114443.

[3] Beeram, D. (2024). Combining deep learning and heuristic search for efficient text summarization. International Research Journal of Engineering and Technology (IRJET), 11(8), 23-34.

[4] Bienvenu, M., Bourgaux, C., & Jean, R. (2024). Cost-based semantics for querying inconsistent weighted knowledge bases. In Proceedings of the 21st international conference on principles of knowledge representation and reasoning (pp. 167-177). Hanoi: CAI Organization. doi: 10.24963/kr.2024/16.

[5] Bourgaux, C., Guimarães, R., Koudijs, R., Lacerda, V., & Ozaki, A. (2024). Knowledge base embeddings: Semantics and theoretical properties. In Proceedings of the 21st international conference on principles of knowledge representation and reasoning (pp. 823-833). Hanoi: International Joint Conferences on Artificial Intelligence Organization. doi: 10.24963/ kr.2024/77.

[6] Gabriel, A. (2020). Kensho derived Wikimedia dataset. Retrieved from https://www.kaggle.com/datasets/ kenshoresearch/kensho-derived-wikimedia-data.

[7] George, S., Elayidom, M.S., & Santhanakrishnan, T. (2019). Semantic desktop search engine using graph database. International Journal of Recent Technology and Engineering, 8(1S2), 373-375.

[8] Gupta, A., & Singh, T. (2024). Study of various frameworks to develop intelligent chatbots. International Journal of Innovative Science and Research Technology (IJISRT), 9(4), 2969-2978. doi: 10.38124/ijisrt/IJISRT24APR1290.

[9] Kaya, C., Kilimci, Z.H., Uysal, M., & Kaya, M. (2024). A review of metaheuristic optimization techniques in text classification. International Journal of Computational and Experimental Science and Engineering, 10(2). doi: 0.22399/ ijcesen.295.

[10] Li, C., Liang, M., & Qiu, D. (2022). An intelligent search system based on knowledge graph. In 2022 International conference on artificial intelligence of things and crowdsensing (AIoTCs) (pp. 66-70). Nicosia: IEEE. doi: 10.1109/ AIoTCs58181.2022.00017.

[11] Lindemann, N.F. (2024). Chatbots, search engines, and the sealing of knowledges. AI & Society. doi: 10.1007/s00146024-01944-w.

[12] Mohabir, S.E., & Joshi, Y.C. (2024). A bibliometric analysis of the knowledge base on multinational corporations’ behavior. SN Business & Economics, 4, article number 105. doi: 10.1007/s43546-024-00705-7.

[13] Morayo, A., Samuel, J., Kennedy, O., Adeyinka, A., Adenugba, A., & Imhade, O. (2024). Development of an artificial intelligent health chatbot for improved telemedicine. In C. So In, N.D. Londhe, N. Bhatt & M. Kitsing (Eds.), Information systems for intelligent systems. ISBM 2023. Smart innovation, systems and technologies (Vol. 379, pp. 585600). Singapore: Springer. doi: 10.1007/978-981-99-8612-5_48.

[14] Rathje, S., Mirea, D.-M., Sucholutsky, I., Marjieh, R., Robertson, C., & Van Bavel, J. (2024). GPT is an effective tool for multilingual psychological text analysis. Proceedings of the National Academy of Sciences of the United States of America, 121, article number e2308950121. doi: 10.1073/pnas.2308950121.

[15] Roy, S., Bharaty, A., Sarkar, S., Sehgal, M., & Panchal, R. (2024). A hybrid ensemble approach for short-text sentiment analysis integrating deep learning and traditional machine learning methods. ResearchGate. doi: 10.13140/ RG.2.2.15182.88643.

[16] Sattar, N.S., & Arifuzzaman, S. (2018). Parallelizing Louvain algorithm: Distributed memory challenges. In 2018 IEEE 16th Intl conf on dependable, autonomic and secure computing, 16th intl conf on pervasive intelligence and computing, 4th intl conf on Big Data intelligence and computing and cyber science and technology congress (DASC/PiCom/DataCom/ CyberSciTech) (pp. 695-701). Athens: IEEE. doi: 10.1109/DASC/PiCom/DataCom/CyberSciTec.2018.00122.

[17] Simian, D., & Șerban, M.-E. (2024). Improving search query accuracy for specialized websites through intelligent text correction and reconstruction models. Information, 15, article number 683. doi: 10.3390/info15110683.

[18] Sutramiani, N., Arthana, I.M.T., Lampung, P.F., Aurelia, S., Fauzi, M., & Darma, I.W.A.S. (2024). The performance comparison of DBSCAN and K-Means clustering for MSMEs grouping based on asset value and turnover. Journal of Information Systems Engineering and Business Intelligence, 10, 13-24. doi: 10.20473/jisebi.10.1.13-24.

[19] Wu, L., Hu, J., Teng, F., Li, T. & Du, S. (2023). Text semantic matching with an enhanced sample building method based on contrastive learning. International Journal of Machine Learning and Cybernetics, 14, 3105-3112. doi: 10.1007/ s13042-023-01823-8.

[20] Yarovyi, A. & Kudriavtsev, D. (2021). Multi-purpose search to determine the context of a text message based on the dictionary data structure. In 2021 IEEE 16th international conference on computer sciences and information technologies (CSIT) (pp. 65-68). Lviv: IEEE. doi: 10.1109/CSIT52700.2021.9648803.

[21] Yuehgoh, F., Djebali, S., & Travers, N. (2024). Leveraging recommendations using a multiplex graph database. International Journal of Web Information Systems, 20(5). doi: 10.1108/IJWIS-05-2024-0137.

[22] Zhang, Y. et al. (2024). A materials terminology knowledge graph automatically constructed from text corpus. Scientific Data, 11, article number 600. doi: 10.1038/s41597-024-03448-0.

[23] Zhao, Y., & Wang, T. (2024). Knowledge base embeddings for a recommendation based on overlapping knowledge and graph learning. Arabian Journal for Science and Engineering. doi: 10.1007/s13369-024-09573-7.