NK-MSVM: A Lossless Recognition Framework for NK-Cola Education Symbol Set Based on Online Handwriting

With the rapid development of information technology, digital transformation has significantly influenced various industries. In the educational sphere, digital transformation aims to reshape the landscape through the comprehensive application of digital technologies, impacting both the medium of delivery and the mode of knowledge acquisition. Nevertheless, several challenges persist, including data isolation, inadequate identification mechanisms, and inefficient data extraction process. In this study, we propose NK-MSVM, a framework for lossless recognition of the NK-Cola education symbol set based on online handwriting. First, we present the NK-Cola education symbol set, specifically designed for K-12 education. Subsequently, we implement classifiers that integrate the generalization capability of Support Vector Machines (SVM) and the parameter allocation advantage of Mixture of Experts (MoE) to enhance symbol recognition. Finally, to mitigate the potential adverse effects of misjudgments on students, we employ a probability boundary with thresholds, ensuring lossless recognition. Experimental results demonstrate that the proposed method effectively achieves lossless recognition of the NK-Cola education symbol set. This research contributes a viable solution to address the challenge of lossless data identification in educational digitalization.