方法证据记录
Text Normalization
Text normalization is an NLP preprocessing pipeline that converts noisy, abbreviated, or misspelled text — such as SMS messages, social-media posts, and OCR output — into a clean, standardised form. It is a prerequisite step for virtually every downstream NLP task, ensuring that inconsistent surface forms do not degrade tokenisation, parsing, or classification. The method gained systematic academic treatment through Baldwin and Li (2015) and Sproat and Jaitly (2017).
源记录
引文逐字复制自方法源记录。这些引文不代表任何层级的验证。
Text Normalization (Noisy-Text Standardisation)
分类方法记录 · process-pipeline / text-mining
- Baldwin, T. & Li, Y. (2015). An In-depth Analysis of the Effect of Text Normalization in Twitter. NAACL-HLT 2015. · URL
- Sproat, R. & Jaitly, N. (2017). RNN Approaches to Text Normalization: A Challenge. arXiv:1611.00068. · URL
精选声明
声明已持久化到证据分类账中，每个声明都有自己的评估。
尚无精选声明
当分类账中没有声明时，此视图不会自行创建声明评估。
相关方法
从方法图中生成，显示为机器建议的关系 — 不推断任何证据声明。