Week 10 · Building a Corpus 语料库建设

Nov 16, 2026 · ← Course home

Planned session. The computational half begins — and it starts with data, not code. We build small corpora from scratch: where to collect texts (and what copyright allows), file formats (txt, csv, xml), encoding pitfalls, cleaning (headers, footers, noise), sentence splitting, and parallel-corpus alignment for Chinese–English material. By the end of the session every group has a working mini-corpus on its own laptop.

Everything is done with graphical tools and plain text editors. What you need is curiosity, not coding experience.

Key concepts: representativeness and sampling; plain text workflow; cleaning; sentence segmentation; alignment.

Slides, templates and a step-by-step handout will be posted here after class.