Leveraging Contextual Sentences for Text Classification by Using a Neural Attention Model

Table 1

Dataset summary statistics.

Dataset	C	V (k)	Train (k)	Validation (k)	Test (k)
SWDA	43	20	193	23	5
MRDA	5	12	78	16	15

C is the number of classes; V is the vocabulary size. The train/validation/test indicates the number of utterances in all dialogs.