SVD Subspace Projections for Term Suggestion Ranking (2004)
by
And Clustering David
,
David Gleich
In Technical Report, Yahoo! Research Labs
Add To MetaCart
Abstract:
In this manuscript, we evaluate the application of the singular value decomposition (SVD) to a search term suggestion system in a pay-for-performance search market. We propose a novel positive and negative relevance feedback method for search refinement based on orthogonal subspace projections. We apply these methods to the subset of Overture's market data and demonstrate a clustering effect of SVD.
Citations
| 1636 | Indexing by latent semantic analysis – Deerwester, Dumais, et al. - 1990 |
| 1439 | Modern Information Retrieval – Baeza-Yates, Ribeiro - 1999 |
| 1241 | Matrix Computations – Golub, Loan - 1993 |
| 183 | Improving the retrieval of information from external sources – Dumais - 1991 |
| 114 | C.: ARPACK user’s guide. Solution of large-scale eigenvalue problems with implicitly restarted Arnoldi methods – Lehoucq, Sorensen, et al. |
| 80 | Understanding Search Engines: Mathematical Modeling and Text Retrieval – Berry, Browne - 1999 |
| 76 | vector spaces, and information retrieval – Berry, Drmac, et al. - 1999 |
| 31 | A unifying theorem for spectral embedding and clustering – Brand, Huang - 2003 |
| 13 | Applications of linear algebra in information retrieval and hypertext analysis – Kleinberg, Tomkins - 1999 |
| 11 | Telcordia LSI engine: Implementation and scalability issues – Chen, Stoffel, et al. - 2001 |
| 9 | Clustering of bipartite advertiser-keyword graph – Carrasco, Fain, et al. - 2003 |

