CPC G06F 40/35 (2020.01) [H04M 3/5141 (2013.01); H04M 3/5191 (2013.01); G06Q 30/0241 (2013.01)] | 20 Claims |
1. A computer-implemented method comprising:
identifying, by one or more processors, a conversation block within a conversation data object;
identifying, by the one or more processors, an initial sentiment score for the conversation block;
identifying, by the one or more processors, a subjectivity probability value for the conversation block that describes an inferred likelihood that the conversation block is characterized by one or more subjective sentiments;
generating, by the one or more processors and a stage-wise penalty factor determination model, a stage-wise penalty factor for the conversation block that describes a stage-aware sentiment score for the conversation block based at least in part on a related conversation stage of a plurality of conversation stages associated with the conversation data object;
generating, by the one or more processors, and based at least in part on the initial sentiment score, the subjectivity probability value, and the stage-wise penalty factor, a regularized sentiment score for the conversation block; and
causing, by the one or more processors, one or more actions to be performed based at least in part on the regularized sentiment score.
|