MACC'97 Session: Information retrieval
It is hard for information agents to integrate heterogeneous semi-structured information, since they use different structures to express the same information. A common ontology or ontologies could resolve this problem. However, it would be very difficult to build such an ontology for a dynamic information space such as the WWW. Thus, automatic reconciliation of differences is needed for open integrative agents. Example-Based Frame Mapping (EBFM) is the technique developed by us for this purpose. EBFM generates a frame map which has equivalent slot pairs of frame structure of the information. However, it is difficult to have a global map in multi-agent system handling a large information space. We propose a distributed EBFM where an agent has a local map and searches other maps to make correspondence with a frame out of his map if needed.
semi-structured information, information retrieval, frame mapping, ontology, EBFM
PS
file(+gzip) (in Japanese)