Abstract
Data structures for similarity search are commonly evaluated on data in vector spaces, but distance-based data structures are also applicable to non-vector spaces with no natural concept of dimensionality. The intrinsic dimensionality statistic of Chávez and Navarro provides a way to compare the performance of similarity indexing and search algorithms across different spaces, and predict the performance of index data structures on non-vector spaces by relating them to equivalent vector spaces. We characterise its asymptotic behaviour, and give experimental results to calibrate these comparisons.
Originalsprog | Engelsk |
---|---|
Titel | Proceedings of the 12th International Conference on String Processing and Information Retrieval (SPIRE 2005), Buenos Aires, Argentina, November 2--4, 2005 |
Antal sider | 12 |
Vol/bind | 3772 |
Forlag | Springer London |
Publikationsdato | 2005 |
Sider | 103-114 |
Status | Udgivet - 2005 |
Udgivet eksternt | Ja |
Emneord
- Similarity search
- Vector spaces
- Distance-based data structures
- Intrinsic dimensionality
- Performance prediction