Information Technology Reference
In-Depth Information
The question therefore arises as to the way in which social and cultural factors may be
linked to non-verbal behavior. In fact, socio-cultural attributes and non-verbal
behavior may continuously influence one another and interact in dynamic and
complex ways.
5.2 Reason for Employing a Bayesian Network
We employed a Bayesian network to represent the model used in the current study as
this is capable of being exploited in a bi-directional way. Therefore, the model can be
used to set or modify the nonverbal behavior of an agent by setting the evidence for a
given culture and social relationship as well as to infer the cultural background and
social situation from given nonverbal behavior. .A further reason for employing a
Bayesian network as the basis for modeling relates to its ability to deal with
incomplete and unpredictable information as well as uncertainties at any stage of
operation. This capability is crucial for present purposes as the link between social,
cultural, and nonverbal behavior involves a many to many mapping procedure.
5.3 The Parameter-Based Socio-Cultural Model
In order to create a Bayesian network that is able to predict non-verbal behavior, the
GeNie [13] modeling system was employed. Fig. 3 illustrates the updated version of
the Bayesian network as employed in the present model. To maintain consistency
with the previous model we had created [1], a social relationship layer was added. In
designing this new model, we carefully investigated the empirical results obtained
from the present analysis with regard to the two social relationships outlined i.e., a
first time encounter and an interaction with someone of a higher status. Based on the
Fig. 3.
Bayesian network model for predicting posture characteristics
