M. W. Berry, T. Do, G. W. O'Brien, V. Krishna, and S. Varadhan. Svdpackc (version 1.0) user's guide. University of Tennessee, April 1993.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....that some inter topic pairs can be close enough to be confused, the average and the standard deviation show that such pairs are extremely rare. Results from experiments with different size parameters are also similar in spirit. In this and the other experiments reported here, we used SVDPACKC [2] for singular value decomposition. Synonymy We end this section with a brief discussion of synonymy in the context of LSI. Let us consider a simple model in which two terms have identical co occurrences (this generalizes synonymy, as it also applies to pairs of terms such as supply demand and ....

M. W. Berry, T. Do, G. W. O'Brien, V. Krishna, and S. Varadhan. SVDPACKC (Version 1.0) User's Guide. University of Tennessee, April 1993.

.... 18 is decomposed all the way in a minute on a small laptop, using the command svd(A,0) in Matlab, while a bigger 70000 Theta 90000 sparse matrix from an information retrieval application can be decomposed until k = 200 in 18 hours on a Sun SS 10 workstation using a Lanczos algorithm from SVDPACK [2]. 3 Estimating data not (yet) observed Now assume that we have observed n variables for m individuals in the data matrix X. Compute its SVD (1) and decide to keep k components, either by cross validation or exhaustion of computer resources, X U k Sigma k;k V T k : 2) The rows of U now give ....

M. W. Berry, SVDPACKC: Version 1.0 user's guide, Tech. Report CS93 -194, University of Tennessee, Knoxville, TN, 1993. Cited in [4].

....analysis, the singular value decomposition not only reduces the dimensionality of the input space, it also yields a collection of term and document vectors that are located in the same low dimensional output space. For the purpose of the present research, we chose the las2 utility of SVDPACKC[4] in order to perform the singular value decomposition (as per the recommendation in [1] We configured las2 to produce a maximum of 300 singular values (it found 182) 2.4: Create vectors for the Internet users in our modeling data set. In order to create LSA vectors for Internet users, we ....

M. W. Berry et al., SVDPACKC: Version 1.0 User's Guide, Tech. Rep. CS-93-194, University of Tennessee, Knoxville, TN, October 1993.

....that some inter topic pairs can be close enough to be confused, the average and the standard deviation show that such pairs are extremely rare. Results from experiments with different size parameters are also similar in spirit. In this and the other experiments reported here, we used SVDPACKC [2] for singular value decomposition. Synonymy We end this section with a brief discussion of synonymy in the context of LSI. Let us consider a simple model in which two terms have identical co occurrences (this generalizes synonymy, as it also applies to pairs of terms such as supply demand and ....

M. W. Berry, T. Do, G. W. O'Brien, V. Krishna, and S. Varadhan. SVDPACKC (Version 1.0) User's Guide. University of Tennessee, April 1993.

....that some inter topic pairs can be close enough to be confused, the average and the standard deviation show that such pairs are extremely rare. Results from experiments with different size parameters are also similar in spirit. In this and the other experiments reported here, we used SVDPACKC [2] for singular value decomposition. Synonymy We end this section with a brief discussion of synonymy in the context of LSI. Let us consider a simple model in which two terms have identical co occurrences (this generalizes synonymy, as it also applies to pairs of terms such as supply demand and ....

M. W. Berry, T. Do, G. W. O'Brien, V. Krishna, and S. Varadhan. SVDPACKC (Version 1.0) User's Guide. University of Tennessee, April 1993.

....(D, T , and Y j Z T j ) discussed in Section 4.1. SVD updating exploits the previous singular values and singular vectors of the original termdocuments matrix A as an alternative to recomputing the SVD of A in Equation (9) In general, the cost of computing the SVD of a sparse matrix [4] can be generally expressed as I Theta cost (G T Gx) trp Theta cost (Gx) where I is the number of iterations required by a Lanczos type procedure [2] to approximate the eigensystem of G T G and trp is the number of accepted singular triplets (i.e. singular values and corresponding left ....

....a sample 3 of about 70; 000 documents and 90; 000 terms was used. Such term by document matrices (A) are quite sparse, containing only :001 :002 non zero entries. Computing A 200 , i.e. the 200 largest singular values and corresponding singular vectors, by a single vector Lanczos algorithm [4] required about 18 hours of CPU time on a SUN SPARCstation 10 workstation. Documents not in the original LSI analysis were folded in as previously described in Section 3.3. That is, the vector for a document is located at the weighted vector sum of its constituent term vectors. Although it is ....

[Article contains additional citation context not shown here]

M. W. BERRY ET AL., SVDPACKC: Version 1.0 User's Guide, Tech. Rep. CS--93--194, University of Tennessee, Knoxville, TN, October 1993. 34

....Assigning Contributed Papers To Meeting Sessions To organize the contributed sessions for the SIAM meeting, we constructed a term by abstract matrix A of order 1149 by 128, corresponding to the 128 submitted papers. A reduced LSI model of rank k = 44 (A 44 ) was generated with SVDPACKC (see [3]) to produce both term and document (abstract) vector representations. Each of the 19 different themes identified in the preliminary announcement of the meeting was used as a query (q) into the vector space model, and the 10 top ranked abstracts (in cosine similarity to q) were recorded. Among the ....

M. Berry, T.Do, G. O'Brien, V. Krishna, and S. Varadhan, SVDPACKC: Version 1.0 User's Guide, Tech. Report CS-93-194, University of Tennessee, Knoxville, October 1993.

....matrices (D, T , and Y j Z T j ) discussed in Section 4.1. SVD updating exploits the previous singular values and singular vectors of the original term documents matrix A as an alternative to recomputing the SVD of A in Equation (9) In general, the cost of computing the SVD of a sparse matrix [3] can be generally expressed as I Theta cost (G T Gx) trp Theta cost (Gx) where I is the number of iterations required by a Lanczos type procedure [2] to approximate the eigensystem of G T G and trp is the number of accepted singular triplets (i.e. singular values and corresponding left ....

....a sample 3 of about 70; 000 documents and 90; 000 terms was used. Such term by document matrices (A) are quite sparse, containing only :001 :002 non zero entries. Computing A200 , i.e. the 200 largest singular values and corresponding singular vectors, by a single vector Lanczos algorithm [3] required about 18 hours of CPU time on a SUN SPARCstation 10 workstation. Documents not in the original LSI analysis were folded in as previously described in Section 3.3. That is, the vector for a document is located at the weighted vector sum of its constituent term vectors. Although it is very ....

[Article contains additional citation context not shown here]

M. W. Berry et al., SVDPACKC: Version 1.0 User's Guide, Tech. Rep. CS--93--194, University of Tennessee, Knoxville, TN, October 1993.

No context found.

M. W. Berry, T. Do, G. W. O'Brien, V. Krishna, and S. Varadhan. Svdpackc (version 1.0) user's guide. University of Tennessee, April 1993.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC