An Algorithm for Robust Relative Influence Values Elicitation (ARRIVE)

This paper proposes an approach to solving node weightings in a tree structure. The tree represents expertise used to quantify risks associated with mental-health problems and it is incorporated within a Web-based decision support system called GRiST. The aim of the algorithm is to find the set of relative node weightings in the tree that helps GRiST simulate the clinical risk judgements given by mental-health experts. The results show that a very large number of nodes (several thousand for GRiST) can have their weights calculated from the clinical judgements associated with a few hundred cases (200 for GRiST). This greatly reduces the experts' elicitation tasks by ensuring they do not need to provide their own estimation of node weights throughout the tree. The approach has the potential for reducing elicitation load in similar knowledge-engineering domains where relative weightings of node siblings are part of the parameter space.