Mysteries Department
Sign inRequest access

Methods

A method is a test that can be run again: what it measures, the code that does it, and every time it has been run. Each run points at the exact data it used and the result it gave.

45methods247runs7kinds of work
Language and textTreats the text as language and measures it the way linguists do.
0
Topic model (LDA) of pages, k by held-out perplexity and fixed k=10
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·NLP (Blei et al. 2003)Inductive·0
18
0
A <-> B alignment of distributional spaces (orthogonal Procrustes)
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·NLP (cross-lingual embedding alignment)Inductive·0
10
0
Unsupervised morphology (Morfessor Baseline)
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·NLP (Creutz & Lagus)Inductive·0
10
0
Distributional word classes (PPMI-SVD embeddings, k-means)
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·NLP / Copiale-style context clusteringInductive·0
5
0
Profile screen against 103 languages
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·Computational linguisticsDeductive·0
3
0
Form-function coupling (spelling vs context similarity)
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·Computational linguisticsDeductive·0
2
0
Held-out context scaling (Kneser-Ney orders 1-8)
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·Language modellingDeductive·0
2
0
List-genre comparison (Latin Library texts)
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·Corpus linguisticsDeductive·0
2
0
Register by locus type (paragraph, label, ring, radial)
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·Corpus linguisticsDeductive·0
2
0
Codex structure and statistics (entropy, Zipf, word length, slot grammar, line effects, A/B keyness, labels)
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·Corpus linguisticsDescriptive·0
0
0
Hidden Markov model of glyph classes
MYS08HTEHX052AX · mystery · click to copyMVoynich Manuscript·Speech / NLPInductive·0
0