Interactive Sinhala Alphabet Learning: A Machine Learning-Based Educational Tool for Preschoolers

This research presents an innovative digital platform designed to assist preschoolers in learning the Sinhala alphabet through interactive letter recognition and writing. The platform integrates the K-Nearest Neighbors algorithm, a non-parametric and instance-based learning method, to classify handwritten Sinhala characters with high accuracy. By training the model on a dataset of 56,000 images collected from children aged 3 to 7, the platform achieves an 89% accuracy rate in character recognition. The tool's real-time feedback system promotes independent learning. The platform was tested with 30 preschoolers and 10 educators to evaluate its effectiveness. Feedback from both groups emphasized its ease of use and suggested adding more interactive and feedback-oriented features for improvement. This study addresses the complexities of the Sinhala script, providing a practical solution for early childhood education, particularly in areas with limited internet access. Furthermore, the use of image processing techniques enhances recognition accuracy across various handwriting styles, enabling the tool to be reliable in various settings. This research enhances the field of early childhood education and machine learning by providing a robust learning aid that blends traditional education with modern technology. It illustrates the potential of this approach in enhancing literacy education through interactive and adaptive digital tools.