Emotionally Intelligent Chatbots: A Systematic Literature Review

Table 7

Manual evaluation criteria.

Evaluation criteria	# of studies	Studies
Emotion accuracy	3	[49, 53, 78]
Response emotion quality and specificity	4	[47, 57, 60, 83]
Response emotion reflection and expression	8	[55–57, 60, 74, 79, 80, 83]
Response emotion diversity	8	[8, 36, 39, 58, 59, 61, 64, 76]
Response emotion appropriateness	10	[8, 36, 39, 48, 58, 59, 61, 64, 69, 82]
Response empathetic emotion intensity	7	[8, 17, 66–68, 82, 84]
Emotion intensity	7	[8, 17, 66–68, 82, 84]
Response grammatical correctness	11	[36, 39, 48, 50, 51, 58, 59, 64, 76, 77, 80]
Response user preference	1	[64]
Response naturalness	4	[7, 50, 54, 80]
Response coherence	5	[7, 39, 54, 68, 80]
Response fluency	4	[17, 49, 68, 80]
Response relevance	13	[17, 36, 48, 49, 52, 57, 61, 66–68, 72, 74, 80, 84]
Response consistency	9	[7, 50–52, 54–56, 73, 84]
Response logic	5	[55–57, 62, 63]
Response intelligible	2	[62, 63]
Response context	1	[72]
Response politeness	1	[72]