DefinePK

Abstract

Dyslexia, a specialized learning condition, affects around 10% of the global population. Adding audio to printed text may produce duplication, but it may be advantageous for kids with dyslexia who need help reading. Studying both the learning process and the learning results in kids with and without dyslexia can shed light on this problem and assist in determining if the redundancy effect is constrained. Most prior electroencephalogram (EEG) tests on people with and without dyslexia identified disparities in the challenges of those with dyslexia. In this study, we provide a model for predicting readers with and without dyslexia based on EEG signals from the brain obtained with BrainSensor equipment. This article treats signals using Empirical Mode Decomposition (EMD) and Singular Spectrum Analysis (SSA). After that, these output signals are given to Deep Forest Classifier to predict dyslexia students. The experiments are carried out on collected signals and validated its performance using four parameters: Accuracy, recall, precision, and F-measure. The proposed model is compared with five existing Machine Learning (ML) and Deep Learning (DL) techniques implemented with SSA-EMD, SSA, and EMD for performance analysis. The proposed Deep Forest Classifier (DFC) model performs better while executing both SSA-EMD and yields 98% accuracy.