Enhancing Preschool Literacy: A Machine Learning Approach to Sinhala Alphabet Recognition

This paper proposes a new tool to help preschool children learn the Sinhala alphabet, as it employs the K-Nearest Neighbors algorithm for handwritten Sinhala character recognition and prediction. The interactive platform also addresses the challenges that arise from conventional classroom learning systems, such as the lack of quick feedback, one-off learning, and inflexibility for individual learners. A crucial aspect of the Sinhala script is that it contains fifty-six characters, making character recognition particularly tedious for preschool children. The K-Nearest Neighbors algorithm appears to fill that gap impressively, as it can determine the handwriting of individuals and convert it into usable pieces for students. Usability testing indicated that preschoolers achieved a defined accuracy threshold of 90%, while the model's accuracy was 87%. These indicators provide internal evidence of the tool's actual utility in classroom interactions. The present study highlights the cultural relevance of HCI applications for children in early years education and advances the fields of machine learning systems and the preservation of the Sinhala language. The outcomes of this research demonstrate the technical correctness of the solution and its potential usefulness in applying this method to other languages and educational systems.