Applied aspects of mathematical application of latent-semantic analysis to classification of scientific texts

Ismukanova Aigerim Nauryzbayevna

The classification of scientific texts in the Russian and Kazakh languages, by means of assignment of a universal decimal code (UDC) is an actual problem nowadays. The problem of the classification of scientific texts is easily solved for the English language due to its simplicity of morphology and syntax. On this way there is a number of unresolved tasks for the Russian language and practically the usage of analogical reception for the Kazakh language is not investigated. For the Russian language several researches of applicability of different approaches were conducted. New technologies for the LSA model could represent a important advance of the assessment of scientific texts.

Keywords:latent semantic analysis (LSA), machine learning (ML), classification.