A Data-Fusion Approach to Representing Personality Traits, Values, Beliefs and Behavior Descriptions

Boon-Kiat Quek, Kayo Sakamoto, Andrew Ortony

Abstract: Starting with currently available data sets that capture the relationships between different pairs of personality-related constructs (for instance, descriptions of behaviors and personality traits), we describe a way of representing personality traits, human values, beliefs, and behaviors in one integrated network-centric model. We propose that such a unified representation can contribute to the making of predictions from one type of construct to another, for example, between personality characteristics and value preferences. Using a spreading activation mechanism, we show how, in principle, such a model can be used to make inferences about the plausibility of unobserved characteristics of an individual, given a very limited initial set of known characteristics.