Towards a Prediction and Data Driven Computational Process Model of Emotion

Abstract

Starting from the assumption that computational models of emotion (CME) should produce tangible and useful results, I focus on their potential role in developing theoretical predictions and comparative testing of different theories. Concretely, I suggest a specific type of CME for this purpose, the prediction and data driven computational process (PDCP) model of emotion, as based on largely shared assumptions about the emotion process and the underlying components. After providing an overview of the wide variety of emotion theories and their suitability for computational modeling, the current version of the component process model (CPM) of emotion is described, including the specific predictions amenable to modeling. I then review the empirical data confirming many of these predictions and thus providing a solid basis for the development of CMEs. On this basis, I outline the skeleton of a realistic PDCP model that should allow testing competing theories and models. Specifically, I propose an incremental approach to concrete static and dynamic modeling efforts using recently developed advanced experimental procedures and assessment instruments as well as expert systems.