This emotion system is meant to be an abstract representation of an agent’s internal emotional state that is influenced with a combination of external input sensors and sentiment appraisal rules. Emotional Cognition is a subfield of cognitive science, with many potential emotional models.
This Emotion Space uses a 2D representation of an emotional plane, with the axis Valence and Confidence. The ranges of these two factors map out the Emotional Space, which creates a mapping of Valence and Confidence values to labeled emotions, shown below. I settled on a 2D emotional graph (omitting the academically popular energy or arousal axis) for simplicity. Ultimately, these 8 labeled emotions capture quite a large range of expressivity, and the system is designed such that adding another axis later on, if we decide it is necessary, would be quite simple. An example of a differentiation that might happen would be Joyful (1, 1, 0) (neutral energy) could become either Ecstatic (1, 1, 1) (high energy) or Satisfied (1, 1, -1) (low energy). As a proof of concept for the emotion system, it was decided these nuances represent far more work than I was comfortable investing without user feedback.
Valence, as used in psychology, especially in discussing emotions, means the intrinsic attractiveness/”good”-ness (positive valence) or averseness/”bad”-ness (negative valence) of an event, object, or situation. The term is also used to characterize and categorize specific emotions. Confidence refers to an emotional sentiments’ reflection of personal self- confidence.
Emotions decay. This emotional baseline is 0.45 valence, 0.2 confidence. The decay rate of this emotional state varies based on the difference between emotional state and baseline emotions. Without any interaction, the state will always return to this state within an hour, despite any original impact.