Abstract:

Information retrieval is an important function which data spaces must provide. This paper describes an index for an information retrieval subsystem of a data space system. Using the relationship between the data in the data space, the relevant data are extracted to compose some basic information units, and then an extended inverted lists based on the basic information units is created for multi_source heterogeneous data. The experiments show that using the index based on basic information units, the system can return comparatively complete semantic information.

Key words: information retrieval;data space;index