Semantic Network Integrity Maintenance via Heuristic Semi-Automatic Tests

In this article we discuss issues connected with maintaining content integrity of general-purpose semantic network that is in development. Construction of a semantic network from scratch is a long process that usually requires both linguistic work done by hand and semi-automatic methods to add or translate the data which must be subsequently reviewed. In this process many systemic and/or language- specific errors may appear in the data over time. We will introduce a method to cope with this issue systematically.