PROBLEM TO BE SOLVED: To improve the detection accuracy of a speaker transition in a discussion. SOLUTION: A discussion analyzer 1 according to an embodiment of the present invention has an information acquisition unit 111 that acquires the utterance amount of each of a plurality of participants in a discussion in which a plurality of participants participate, and a first in the discussion. Multiple participants based on the maximum speaker identification unit 112 that identifies the maximum speaker with the maximum utterance volume among the plurality of participants in each time range, and the change in the maximum speaker in the first time range. It has an output unit 116 that outputs transition information indicating the transition of the speaker generated between the persons. [Selection diagram] Fig. 2
展开▼