We present the first public, user-friendly system for Galician poetry scansion, a symbolic system derived from a well-performing mixed-meter Spanish scansion library. We adapted its resources to Galician and added a preprocessing module. The system achieves 88% per-line accuracy in exact stress-pattern match on data unseen during development, and has practical value: First, it helps create a large annotated corpus to train scansion systems. Second, its web interface can help engage a non-specialist public. Third, its current accuracy is helpful for annotating large volumes of poetry and studying metrical trends in Computational Literary Studies use cases.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。