MATHEMATICAL MODELS FOR DETERMINING THE SEMANTIC SIMILARITY OF TEXTS

Allaberganova, Nasiba

Al-Farg'oniy avlodlari · 2026-yil

Annotatsiya

The article investigates the problem of determining semantic similarity of Uzbek texts using mathematical modeling. The study aims to integrate semantic analysis, vector space representation, similarity calculation, and decision-making based on a threshold value into a unified model. The methodology includes text preprocessing, semantic vector generation, and similarity measurement using cosine similarity. Experiments were conducted in the Python environment, and the results were evaluated using Accuracy, Precision, Recall, and F1 metrics. The findings show that vector-based approaches represent semantic relationships more effectively than lexical-statistical models, while threshold optimization reduces misclassification and improves the overall accuracy of text similarity detection systems.

Maqola ma’lumotlari
MualliflarAllaberganova, Nasiba
JurnalAl-Farg'oniy avlodlari
Nashr sanasi2026-06-07
Son2
Betlar282-286
TilRus

Kalit so‘zlar

Semantik analiz, matn o‘xshashligi, matematik model, NLP, embedding, vektor fazo modeli, cosine similarity, semantik masofa, Chegaraviy qiymat modeli., Semantic analysis, text similarity, mathematical model, NLP, embedding, vector space model, cosine similarity, semantic distance, Boundary value model.

Ilmiy soha

Al-Farg'oniy avlodlari jurnalidan boshqa maqolalar

Al-Farg'oniy avlodlari — barcha maqolalar