Title :
Resolution spaces: a topological approach to similarity
Author :
Georgatos, Konstantinos
Author_Institution :
Dept. of Math., City Univ. of New York, NY, USA
Abstract :
A central concept for information retrieval is that of similarity. Although an information retrieval system is expected to return a set of documents most relevant to the query word(s), it is often described as returning a set of documents most similar to the query. The authors argue that in order to reason with similarity we need to model the concept of discriminating power. They offer a simple topological notion called resolution space that provides a rich mathematical framework for reasoning with limited discriminating power, avoiding the vagueness paradox
Keywords :
database theory; indexing; inference mechanisms; information retrieval; topology; discriminating power; documents; information retrieval; mathematical framework; query word; reasoning; resolution spaces; similarity; topological approach; vagueness paradox; Educational institutions; Electronic mail; Frequency; Indexing; Information retrieval; Mathematical model; Mathematics; Testing; Topology; Utility theory;
Conference_Titel :
Database and Expert Systems Applications, 2000. Proceedings. 11th International Workshop on
Conference_Location :
London
Print_ISBN :
0-7695-0680-1
DOI :
10.1109/DEXA.2000.875080