Table 1.

Features for machine learning classification models.

Task	Feature groups and number of features (n) in each group
Picture description	Cookie Theft image information units (13)  Part-of-speech (15), context-free-grammar rules (44), syntactic complexity (24), vocabulary richness (4), psycholinguistic (5), repetitiveness (5)
Reading	Syllable count (1), pause count (1)a, total duration (1), total time spent speaking (1), proportion of time spent speaking (1), speech rate (1), average syllable duration (1), pauses per syllable (1)a, pause rate (1)a, pause duration (3)a
Experience description	Part-of-speech (15), context-free-grammar rules (44), syntactic complexity (24), vocabulary richness (4), psycholinguistic (5), repetitiveness (5)

^aThese features were computed using acoustic data and transcript data and are also affected by method of pause detection (ie, acoustic vs text data).