Description:

The objective of this research is to develop a new document organization technique, which simulates human-like processing of the text information (we focus on the problem of multiple levels of abstraction occurring in our natural language). The proposed approach is driven by recent discoveries in the areas of (1) graph-based data mining, (2) hierarchical clustering, and (3) ontology-based data processing. We want to develop mechanisms allowing for a human-like, sense-based querying rather than to continue development of searches dependent solely on the frequency of matching terms. This should allow a broad range of computer users, who are not experts in the searched topics, to have similar searching capabilities, as the experts have.