Modification of an automatic method for extracting causal relationships based on templates and a Bayesian classifier
https://doi.org/10.21822/2073-6185-2025-52-1-162-172
Abstract
Objective. The aim of the study is to modify the automatic method for extracting cause-and-effect relationships.
Method. The study is based on the original method of Antonie Sorgente with its subsequent modification.
Result. A method for extracting cause-and-effect relationships is proposed. The method involves the combined use of statistical data and machine methods. The original method was modified by translating the method to modern libraries such as NLTK and Spacy. The rules formed by the author were reworked and added to the Dependency Matcher module of the Spacy library. The number of keywords for each rule was increased. The method also takes into account synonyms, calculates Bayesian statistics and smooths Laplace for zero probabilities. Based on the difference in data with and without PSS, a multiplier coefficient was introduced to compensate for the skew of classes in the data.
Conclusion. The developed method was tested on the original data of the original method and showed improved metrics relative to the original method on training and test data.
About the Authors
Н. В. Shtanchaev
Daghestan State Technical University
Russian Federation
Russian Federation
Hairutin B. Shtanchaev, Cand. Sci. (Eng), Assoc. Prof., Department of Software for Computer Engineering and Automated Systems, Senior Engineer of APCS,
170 I. Shamil Ave., Makhachkala 367015
Z. T. Mugutdinov
OOO "Salavatsteklo Caspian"
Russian Federation
Russian Federation
Zalibek T. Mugutdinov, Postgraduate Student, Department of Software for Computer Engineering and Automated Systems,
1 Zavodskaya St., Republic of Daghestan, s. Korkmaskala 368080
References
1. Shtanchaev H.B. Non-statistical methods of automatic extracting causal relationship from text// H.B. Shtanchaev. Izvestiya YuFU. Tekhnicheskie nauki. 2023;2(232):273-280. – DOI 10.18522/2311-3103- 2023-2-273-280. – EDN JZUBSO.1. (In Russ)
2. Shtanchaev, H.B. Statistical methods of automatic extracting causal relationship from text(review)// H.B. Shtanchaev. Izvestiya YuFU. Tekhnicheskie nauki. 2023;6(236):105-114. – DOI 10.18522/2311-3103-2023-6-105-114. – EDN TBHUWW. (In Russ)
3. Fellbaum, Christiane (2005). WordNet and wordnets. In: Brown, Keith et al. (eds.), Encyclopedia of Language and Linguistics, Second Edition, Oxford: Elsevier, 665-670.
4. VerbNet. A Computational Lexical Resourse for Verbs. Internet-resurs. Sposob dostupa - https://verbs.colorado.edu/verbnet/
5. Text Rettieval Conference(TREC). Internet-resurs. Sposob dostupa - https://en.wikipedia.org/wiki/Text_Retrieval_Conference
6. G.V. a. F.M. Antonio Sorgente. Automatic extraction of cause-effect relations in. Institute of Cybernetics “Eduardo Caianiello” of the National Research Council, Yanvar' 2013.
7. L.A. Dalton, E.R. Dougherty. Optimal Bayesian Classification, SPIE--The International Society for Optical Engineering, 2020:363p.
8. Sayed, A.H. Inference and Learning from Data: Learning. Cambridge: Cambridge University Press. – Chapter 55, 2022:2341-2356
9. Ananda P. Noto, Dewi R.S. Saputro; Classification data mining with Laplacian Smoothing on Naïve Bayes method. AIP Conf. Proc. 28 November 2022; 2566 (1): 030004.
10. Dependency Matcher Internet-resurs. Sposob dostupa - https://spacy.io/usage/rule-basedmatching#dependencymatcher.
11. SpaCy 101. Vse chto nam nuzhno znat'. Internet-resurs. Sposob dostupa - https://webdevblog.ru/spacy-101- vse-chto-vam-nuzhno-znat-chast-1
12. Metrics of binary classification. Internet.Link - https://loginom.ru/blog/classification-quality (In Russ)
Review
For citations:
Shtanchaev Н.В., Mugutdinov Z.T. Modification of an automatic method for extracting causal relationships based on templates and a Bayesian classifier. Herald of Dagestan State Technical University. Technical Sciences. 2025;52(1):162-172. (In Russ.) https://doi.org/10.21822/2073-6185-2025-52-1-162-172
Views:
63
Email this article 
