Automatic pattern detection in stakeholder networks

The values and opinions of the stakeholders involved in a decision-making process are the key to its outcome. Reflection on how stakeholders perceive a situation, and on the consequences of these perceptions for the decision outcome is an intellectually demanding exercise. To support analysis, we have developed a conceptual modeling tool called DANA (Dynamic Actor Network Analysis). The modeling language is based on the policy network paradigm and embodies concepts from cognitive mapping and linguistic approaches to approximate reasoning. In this paper, we investigate how certain interesting properties of stakeholder networks modeled with DANA can be determined algorithmically. Automatic detection of the factors most relevant to a policy situation, and of disagreement and conflict among stakeholders may help the analyst in focusing her analysis.