A Comprehensive Survey on Predicting Dyslexia and ADHD Using Machine Learning Approaches

Abstract

AbstractAttention deficit hyperactivity disorder (ADHD) and dyslexia are neurological disorders characterized by vague comprehension and generally refer to poor reading and writing ability. It influences some specific populations, i.e. school-aged children, specifically male children. Therefore, it leads to risks and consequences like low self-esteem and poor academic performance for the entire lifetime. The long-term need of the researchers is to model appropriate ADHD and dyslexia prediction approaches to help the affected children community. Based on this, various machine learning (ML) approaches are implemented using multiple online available datasets with attaining better prediction performance and classification accuracy. Moreover, acquiring the clinical acceptability with these existing approaches provides specific research challenges like dataset privacy, appropriate classifier model, good optimizer and feature selection, hyperparameter selection and over-fitting problems. This work provides an extensive review of various existing ML approaches, image processing approaches, analysing multiple performance metrics, research gaps, and research challenges. All these processes attempt to perform critical analysis towards the provided ML which comes for predicting ADHD and dyslexia and makes an appropriate way to the users of ML approaches to facilitate the model performance is in an acceptable range. Therefore, researchers can envisage higher prediction and classification performance with the available clinical relevances using ML which comes for ADHD and dyslexia prediction by addressing the potential research challenges.KeywordsADHDDyslexiaMachine learningPredictionResearch challenges