Continued work towards creation of a web-based platform that supports organic growth of an expert knowledge network at the University of Tokyo and associated research institutions for sustainability science is described. The network is intended to provide value-adding services in the context of sustainability issues that meet specific information needs of a wide range of users from experts in sustainability science to concerned individuals in society with no specific scientific training. Here, we present an assessment of the semantic matching accuracy between a subset of the EKOSS semantic statements that we have authored for AGS research projects over the past several years. Using an algorithm that we have developed for calculating the semantic similarity of the graph representations of pairs of semantic statements, we construct a similarity matrix between the projects. We also construct a similarity matrix using a na?ve matching technique that does not utilize logic and rule-based inference against the semantic relationships stated in the semantic statements. Comparison of the degrees of semantic similarity calculated with and without the use of inference show that the semantic similarity calculation can indeed find matches between semantic statements that mean the same thing even if they do not say the same thing explicitly.
Expert knowledge, knowledge network, logical inference, semantic matching, knowledge description, sustainability science