Spreading activation simulation with semantic network skeletons
Spreading activation algorithms are a well-known tool to determine the mutual relevance of nodes in a semantic network. Although often used, the configuration of a spreading activation algorithm is usually problem-specific and experiencedriven. However, an excessive exploration of spreading behavior is often not applicable due to the size of most semantic networks. A semantic network skeleton provides a comprised summary of a semantic network for better understanding the network’s structural characteristics. In this article, we present an approach for spreading activation simulation of semantic networks utilizing their semantic network skeletons. We show how expected spreading activation behavior can be estimated and how the results allow for further effect detection. The appropriateness of the simulation results as well as time-related advantages are demonstrated in a case study.
Published in: International Journal on Advances in Intelligent Systems, IARIA