Information Technology Reference
In-Depth Information
PAD Temp erament Mo de l .
This model[7] is a very general descriptive sys-
tem for the study of temperament and personality. It is based on same three
dimensions of the PAD Emotion Model (
P-A-D
).
Temperament
is distinguished from emotional states in that it refers to an
individual's stable or lasting emotional characteristics (i.e., emotional traits or
emotional predispositions). More precisely, temperament is an average of a per-
son's emotional states across a representative variety of life situations. A set
of three PAD temperament scales has been developed and shown to provide a
reasonably general description of emotional traits or temperament.
2.3 PAD Space for Emotions and Temperament
The work of Mehrabian postulates that the PAD Space is suitable for represent-
ing emotions of a concrete and isolated event given at a specific instant, and
emotional states (moods) as representing emotional information gathered along
a period of time, thus more stable and persistent in time.
These two separated elements are relevant for the emotional behaviour of
a person. The temperament of a person creates a tendency of the emotional
states that must be achieved if the emotional stimuli are weak or inexistent.
On the other hand, the emotions prompted by the consequences of the events
perceived by a person are relevant by two aspects: the reactive behaviour derived
by a particular strong emotion, and the aggregation of several emotions which
change the emotional state (mood) of a person.
The present work, as we show in the Section 3, introduces this formalization
and a mechanism to translate emotions and moods in the same space.
2.4 Computational Models
Computational models for emotional agents and emotional behaviour take differ-
ent ways to model the dynamics of emotions and their influence on the behaviour
of the agents. In the following, we focus on work that uses OCC as basis for the
analysis and/or synthesis of the emotions, and the models that use PAD as
projection/management system.
In FAtiMA [2], the OCC model is treated as initially conceived: each tag
separates the type of emotion, and each emotion has associated an intensity
derived from the concrete parameters attached to it (for example the desirability
of a consequence can produce
). Then, the emotions produced
are evaluated against a set of rules trying to match some preconditions to trigger
that rule and the actions derived by it.
Other approaches, like WASABI [1] or EBDI [5], use a projection of OCC
emotions into a three dimensional PAD space. They consider the different emo-
tions as a vectorial value of
Joy
or
Distress
3
. In the case of WASABI the emotions are treated
differently depending on whether they are primary or secondary: while the for-
mer are denoted by points, the latter are treated as regions of the space. Once
an event is analysed, the agent is aware of which emotions are more likely to
elicit according to the distance to the central point of each of the emotions, make
R
