MIT Press

Table 2

Data set specifications after preprocessing.

		WALS	Autotyp
L	(# of languages)	2,607	1,063
K	(# of linguistic parameters)	50 or 100
M	(# of model parameters of W, $\tilde{Θ}$ ⁠, and Θ)	760	958
N	(# of discrete features)	152	372

# of binary features		14	229
# of categorical features		138	118
# of count features		0	25

Proportion of items present in the language–feature matrix (%)		19.98	21.54

# of phylogenetic neighbors on average		30.77	7.43
# of spatial neighbors on average		89.10	38.53

		WALS	Autotyp
L	(# of languages)	2,607	1,063
K	(# of linguistic parameters)	50 or 100
M	(# of model parameters of W, $\tilde{Θ}$ ⁠, and Θ)	760	958
N	(# of discrete features)	152	372

# of binary features		14	229
# of categorical features		138	118
# of count features		0	25

Proportion of items present in the language–feature matrix (%)		19.98	21.54

# of phylogenetic neighbors on average		30.77	7.43
# of spatial neighbors on average		89.10	38.53