Acquiring Knowledge from Group Discussion Process Based on Graph Analysis

Currently proposed group decision making approaches, such as expected utility theory, the analytic hierarchy process and fuzzy linguistic approach, require expert working on problem individually, which discourages the cooperation and communication among experts. In this paper, we propose an approach which aids group decision making by analyzing the process of group discussion. Our approach consists of two steps. In the first step, the group discussion process is modeled as directed and weighted graph and in the second step novel algorithms are proposed to evaluate statement quality, assess expert contribution and identify sub-groups approving or disapproving a certain statement respectively.