A Novel Metric for Information Retrieval in Semantic Networks

We propose a novel graph metric for semantic entity-relationship networks for solving two tasks. First, given a semantic entity-relationship graph such as for example DBpedia we find relevant neighbors for a given query node. Second, we search for paths between two given nodes in order to discover interesting links between the nodes. Compared to using the default step metric our approach yields more specific and informative results, as we demonstrate for two semantic web datasets. Moreover, we show that our proposed metric can intuitively be interpreted in terms of random walks. The distances are defined via paths that maximize the log-likelihood of a restricted round trip in such a random process. This also yields a link to the commute distance, which is highly plausible for the described tasks but prohibitively expensive to compute. In comparison, our metric can be calculated efficiently using standard graph algorithms, rendering the approach feasible also for the very large graphs of the semantic web.