Diachronic lexical drift measurement

For a given term and a dated corpus, it quantifies how the vocabulary surrounding that term moves over time.

1. 01

   Distinctive terms per period

   Weighted log-odds with an informative Dirichlet prior (Monroe, Colaresi & Quinn, 2008). Per-term z-score and significance threshold.

2. 02

   Drift magnitude between periods

   Jensen-Shannon divergence, significance assessed via a permutation test.

3. 03

   Full provenance

   Every measure traces back to the source documents. An under-sampled period is flagged, never filled in.

Corpus-agnostic and source-agnostic: parliamentary debates, regulatory texts, press, transcripts, any dated corpus.

No public demo yet, on purpose. Leave your email to hear when it ships.

No spam. One message, the day it is usable.