A SINGLE-LAYER FEEDFORWARD NEURAL NETWORK MODEL FOR MORPHOLOGICAL STEMMING OF UZBEK WORDS

Abstract

This paper presents a single-layer feedforward neural network for morphological stemming of Uzbek words. Each word is encoded using character-level one-hot vectors (810-dimensional input: 27 characters × 30 positions). The model uses sigmoid activation, MSE loss, and gradient descent backpropagation. A systematic hyperparameter search over epochs E ∈ {20,30,40}, learning rates η ∈ {0.5,0.6}, and thresholds θ ∈ {0.70–0.85} across 24 configurations on 3,441 word–stem pairs yields a peak accuracy of 98.34% (E=40, η=0.6, θ=0.70). The results demonstrate that minimal neural architectures can effectively model agglutinative morphology in under-resourced Turkic languages.

More by Ulugbek Salaev

• Comparative Analysis of Neural Sequence Labeling Approaches for Uzbek Morphological ModelingInternational Siberian workshops and tutorials on electron devices and materials · 2026
• Neural Sequence Models for Uzbek Morphological StemmingInternational Scientific-Technical Conference on Actual Problems of Electronic Instrument Engineering · 2025 · Citations: 7
• A Morphology-Aware Neural Network Model for Uzbek Text ClassificationInternational Scientific-Technical Conference on Actual Problems of Electronic Instrument Engineering · 2025 · Citations: 1
• A Feedforward Neural Model for Morphological Stemming of Uzbek WordsInternational Conference on Computer Science and Engineering · 2025 · Citations: 3