Fuzzy Neural Network Models for Early Detection and Support of Learning Disabilities

Abstract

Early identification of learning disabilities (LD) is critical yet challenging due to gradual onset and behavioral variability. We propose a hybrid Fuzzy Neural Network (FNN) that integrates fuzzy feature extraction with neural classification to detect LD indicators in children aged 7–9. Trained on multilingual classroom datasets (N=400), the FNN achieved 92.3% accuracy, outperforming standard neural nets by 7%. In intervention simulation, the model reduced time-to-intervention by 30%. These results suggest scalable, data-driven support for personalized learning support within educational settings.