ILM is a diachronic lexical resource for Malayalam, built by extracting and
cleaning entries from historical Malayalam dictionaries rather than from a
contemporary corpus. It was built to support the Cognate Availability
Principle (CAP): distinguishing culturally-loaded source terms that
translation practice has preserved from routine vocabulary, and supplying
the gloss material needed to preserve them in machine translation.
The… See the full description on the dataset page:
https://huggingface.co/datasets/insightpublica/ilm.
