Detection of Potential Specific Learning Disabilities in Children through Handwriting Analysis Using Machine Learning

Specific learning disability refers to a variety of learning difficulties that impact an individual's capacity to acquire and apply specific academic skills, including reading, writing, mathematics, and reasoning. Learning difficulties, if not identified, can adversely affect children's overall learning experience. Early identification of learning difficulties in children is crucial for providing the necessary support and interventions to help them overcome the challenges. Conventional methods used by psychologists to diagnose learning disabilities include subjective assessments that also examine the behavioral aspects of children. However, these tests are time-consuming and im-practical for large groups of children. Therefore, an automated initial screening method can provide a cost-effective solution for identifying children at risk of learning difficulties who need further testing by specialists. In this study, we present an efficient, low-cost system for identifying children at risk by analyzing the static features of their handwritten text. We extracted six hand-writing features from data collected from typically developing children and those with clinically diagnosed learning disability. These features were used as inputs to a machine learning model to predict the risk of specific learning disability. We compared the performance of various classifiers, and the random forest classifier achieved the best results, with an accuracy of 87.3 %. These findings indicate that the model effectively detects risk in children suggesting its potential as a valuable tool for early identification of specific learning disabilities.