CONSTRUCTION OF THE STRUCTURE OF HOMOGENEITY OF A. FIRDOUSI “SHAHNAME” POEMS ON THE BASIS OF BIGRAMS

Ilhomjon Qodirov

Authors

 

Kosimov A.A. – Candidate of Technical Sciences, Senior Lecturer, Department of Automated Control Systems, Tajik Technical University, Dushanbe, Republic of Tajikistan, abdunabi_kbtut@mail.ru

Kurbonov N.M. – PhD student, Department of Information Technology and Information Protection, Tajik Technical University, Dushanbe, Republic of Tajikistan, nurullo94@gmail.com

Murodov Kh.M. – magistrant, Department of Automated Control Systems, Tajik Technical University, Dushanbe, Republic of Tajikistan

Zulfov Y.O. – magistrant, Department of Automated Control Systems, Tajik Technical University, Dushanbe, Republic of Tajikistan

 

Annotation

 

   The article considers the construction of the homogeneity structure of the poems of A. Firdousi’s work “Shahnameh” on the basis of digrams. The Introduction section and 63 poems of the work are compared with digital portraits based on the distribution of particulars of bigrams of the Cyrillic alphabet of the Tajik language in them. An agglomerative hierarchical classification algorithm is used. The discrete random number classifier method is adopted as the distance between objects. The mathematical model of the classifier was presented as a triad. Its first component is a digital portrait of the text – the distribution of the frequency of letter digrams in the text; the second component is a formula for calculating the distances between a digital portrait and texts, and the third is a machine learning algorithm that implements the hypothesis of “homogeneity” of works written in one language and “heterogeneity” of works written in different languages. Adjustment of the algorithm that uses the table of pairwise distances between all products of the model collection consisted in determining the optimal value of the real parameter for which the error of violating the “homogeneity” hypothesis is minimized. Using the method of the nearest neighbor by the distance matrix, hierarchical clustering of the constituent parts of the product is carried out. When carrying out cluster analysis, according to the “near neighbor” principle, several clusters were obtained, the distance between the clusters is different. The results of the hierarchical classification of objects are presented as a dendrogram.

 

Keywords

 

 Firdousi, Shahnameh, bigram, frequency, distance, classifier, nearest neighbor

 

References

 

  1. Vorontsov K.V. Mathematical methods of teaching by precedents, p. 141, [Electronic resource] – Access mode. – URL: http://www.ccas.ru/voron (accessed 10.03.2021).
  2. Usmanov Z.D. Algorithm for tuning a clusterer of discrete random variables. – DAN RT, 2017, v.60, No. 9, p. 392-397.
  3. Usmanov Z.D. Classifier of discrete random variables. – DAN RT, 2017, v.60, No. 7-8, p. 291-300.
  4. Usmanov Z.D. On a Generalization of the Golden Ratio Formula. – Reports of the Academy of Sciences of the Republic of Tajikistan, 2014, v.57, No. 1, p. 5-8.
  5. Usmanov Z.D., Kosimov A.A. Digital image of “Shahnameh” (“Book of Kings”) by A. Firdowsi. – Reports of the Academy of Sciences of the Republic of Tajikistan, 2014, v.57, No. 6, p. 471-476.
  6. Usmanov Z.D., Kosimov A.A. On the correlation of word forms and word usages in the work of A. Firdowsi “Shahnameh”. – Reports of the Academy of Sciences of the Republic of Tajikistan, 2015, v.58, No. 8, p. 678-683.
  7. Usmanov Z.D., Kosimov A.A. On the question of the position of the culmination point in works of art. – Materials of the 17th scientific-practical seminar “New information technologies in automated systems”. – M., 2014, p. 392-395.
  8. Firdavsi A. – Shokhnoma. – Dushanbe: Adib. – 2007/2008/2009/2010. – Gild 1-10. – 4736 p.
  9. Firdavsi A. Shokhnoma. – Dushanbe: Adib, 2007/2008.
  10. Khudoiberdiev Kh.A., Kosimov A.A. On the correlation of word forms and word usages in the Russian translation of A. Firdousi’s work “Shahnameh”. – Reports of the Academy of Sciences of the Republic of Tajikistan, 2015, v.58, No. 9, p. 786-792.

 

Publication date

2023-10-27

Leave a Comment