Eurasian Journal of Mathematical Theory and Computer Sciences 5-jild 9-son (2025) · 7–12-betlar

METHODS OF SIZE REDUCTION TO INCREASE THE EFFICIENCY OF PERSONAL IDENTIFICATION BY VOICE

Nurimov, Paraxat

DOI: 10.5281/zenodo.17085406 · Manbada o'qish → · PDF (manba serverida)

Annotatsiya

This paper addresses the problem of efficient speaker recognition on resource-constrained devices. The focus is placed on reducing memory and computational costs while preserving the discriminative power of the feature set. To achieve this, dimensionality reduction techniques such as Principal Component Analysis (PCA), Independent Component Analysis (ICA), Linear Discriminant Analysis (LDA), and the Genetic Algorithm (GA) were applied. Experimental results demonstrate that these approaches significantly reduce memory usage and computational complexity while maintaining high recognition accuracy. The proposed methodology is particularly suitable for mobile devices and real-time systems with limited resources

Speaker recognition, dimensionality reduction, PCA, ICA, LDA, Genetic Algorithm, resource-constrained systems, real-time applications.Shaxsni ovozi orqali identifikatsiyalash, o‘lchamni qisqartirish, PCA, ICA, LDA, Genetik algoritm, resurs cheklangan tizimlar, real vaqt ilovalari.

Metadata manbasi: jurnal OAI-PMH arxivi · Sindex to'liq matnni saqlamaydi, manbaga havola beradi.