Reasoning about attitudes of complaining customers

The paper introduces a knowledge engineering methodology to handle scenarios of inter-human interaction. We focus on the attitudes of the conflicting agents, building the formal framework for the representation of these scenarios as sequences of mental actions of the agents. Developed framework facilitates a domain-independent comparative analysis of the textual scenarios of inter-human interactions, as well as a machine learning technique for estimation of the attitudes of involved agents. The formal model for a special case of conflict (customer complaint) is built and the developed framework is applied to justify the scenario as being a valid (consistent, plausible) or invalid complaint. The result of this study is that complaint validity assessment is possible, taking into account the information about mental actions of involved agents only. A formal representation of the mental world of complainants is constructed, where the mental actions are defined in terms of their beliefs and intentions. 2006 Elsevier B.V. All rights reserved. T