DIDACTIC POTENTIAL OF DIGITAL TECHNOLOGIES IN THE DEVELOPMENT OF MATHEMATICAL SPEECH

Abstract

This article examines the didactic potential of digital technologies in bilingual teaching of mathematics and mathematical speech. It demonstrates how subject-language integration, combined with digital technologies, supports the development of mathematical speech and improves academic outcomes. The results indicate that language is a strong predictor of mathematical achievement; support vector machines demonstrated the best results, while ensemble trees demonstrated comparable accuracy with good interpretability of feature importance. A neural network performed worse on a small dataset. Practically, combining digital platforms with lean machine-learning models enables early identification of at-risk students and targeted assistance in bilingual settings. These tools complement teacher decisions by supplying data on each student’s subject and language needs.