MIT Press

Table 2.

CDS F-score minus ADS F-score (in percentages) by algorithm, type of match, and whether human (H) or automatic (A) utterance boundaries were considered.

Algo	H: full	H: UM	H: WM	A: full	A: UM	A: WM	Median
base-utt	2.9	1.4	1.5	3.3	1.8	1.9	1.85
base-syll	1.3	−0.2	0	1.2	−0.2	−0.3	−0.1
phonotactic-unsupervised	−1.6	−2.5	−2.5	−1.3	−2.3	−2.4	−2.35
phonotactic-gold	3	2.6	2.8	3.1	2.8	2.8	2.8
TP-relative	5	0.7	0.9	5	0.4	0.7	0.8
TP-average	−1.6	−2.9	−2.8	−1.5	−2.9	−3.1	−2.85
lexical-incremental	7.9	−0.6	1.6	7.1	1.2	2.3	1.95
lexical-unigram	3.8	2.9	3.6	2.8	2.3	2.4	2.85
lexical-multigram	3.4	1.9	0.2	3.1	−0.8	0.4	1.15
Median	3	0.7	0.9	3.1	0.4	0.7	1.15

Algo	H: full	H: UM	H: WM	A: full	A: UM	A: WM	Median
base-utt	2.9	1.4	1.5	3.3	1.8	1.9	1.85
base-syll	1.3	−0.2	0	1.2	−0.2	−0.3	−0.1
phonotactic-unsupervised	−1.6	−2.5	−2.5	−1.3	−2.3	−2.4	−2.35
phonotactic-gold	3	2.6	2.8	3.1	2.8	2.8	2.8
TP-relative	5	0.7	0.9	5	0.4	0.7	0.8
TP-average	−1.6	−2.9	−2.8	−1.5	−2.9	−3.1	−2.85
lexical-incremental	7.9	−0.6	1.6	7.1	1.2	2.3	1.95
lexical-unigram	3.8	2.9	3.6	2.8	2.3	2.4	2.85
lexical-multigram	3.4	1.9	0.2	3.1	−0.8	0.4	1.15
Median	3	0.7	0.9	3.1	0.4	0.7	1.15

Note. Full means the full child-directed speech (CDS) corpus was used; UM = utterance match: CDS corpus shortened to have as many utterances as the adult-directed speech (ADS) corpus; WM = word match: idem for words.

View Large

This Feature Is Available To Subscribers Only