标记化与增强：基于惯性测量单元的在线手写识别中的写作者差异系统研究

Inertial measurement unit-based online handwriting recognition enables the recognition of input signals collected across different writing surfaces but faces challenges due to uneven character distributions and inter-writer variability. This study systematically investigates two strategies to address these issues: sub-word tokenization and concatenation-based data augmentation. Experiments conducted on the OnHW-Words500 dataset reveal a clear distinction in handling inter-writer and intra-writer variance. On the writer-independent split, structural abstraction through Bigram tokenization significantly enhances performance on unseen writing styles, reducing the word error rate (WER) from 15.40% to 12.99%. In contrast, results on the writer-dependent split show different outcomes.