Open chirila opened 5 years ago
How can we tell whether we are finding orthographic neighbors vs semantic/morphological neighbors?
what is the correlation between orthographic distance [Levenshtein distance] and word embedding vector distance?
produce matrix of orthographic distances, matrix of fastText distance, compare matrices (Mantel test of partial matrix correlation)
cluster/produce trees; use quartets method to assess congruity (expect congruity for morphological neighbors)
[subsumes #1]
How can we tell whether we are finding orthographic neighbors vs semantic/morphological neighbors?
what is the correlation between orthographic distance [Levenshtein distance] and word embedding vector distance?
produce matrix of orthographic distances, matrix of fastText distance, compare matrices (Mantel test of partial matrix correlation)
cluster/produce trees; use quartets method to assess congruity (expect congruity for morphological neighbors)
[subsumes #1]