Algorithm 971: An Implementation of a Randomized Algorithm for Principal Component Analysis
Recent years have witnessed intense development of randomized methods for low-rank approximation. These methods target principal component analysis and the calculation of truncated singular value decompositions. The present article presents an essentially black-box, foolproof implementation for Math...
Uloženo v:
| Vydáno v: | ACM transactions on mathematical software Ročník 43; číslo 3 |
|---|---|
| Hlavní autoři: | , , , , , |
| Médium: | Journal Article |
| Jazyk: | angličtina |
| Vydáno: |
United States
01.01.2017
|
| Témata: | |
| ISSN: | 0098-3500 |
| On-line přístup: | Zjistit podrobnosti o přístupu |
| Tagy: |
Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
|
| Abstract | Recent years have witnessed intense development of randomized methods for low-rank approximation. These methods target principal component analysis and the calculation of truncated singular value decompositions. The present article presents an essentially black-box, foolproof implementation for Mathworks' MATLAB, a popular software platform for numerical computation. As illustrated via several tests, the randomized algorithms for low-rank approximation outperform or at least match the classical deterministic techniques (such as Lanczos iterations run to convergence) in basically all respects: accuracy, computational efficiency (both speed and memory usage), ease-of-use, parallelizability, and reliability. However, the classical procedures remain the methods of choice for estimating spectral norms and are far superior for calculating the least singular values and corresponding singular vectors (or singular subspaces). |
|---|---|
| AbstractList | Recent years have witnessed intense development of randomized methods for low-rank approximation. These methods target principal component analysis and the calculation of truncated singular value decompositions. The present article presents an essentially black-box, foolproof implementation for Mathworks' MATLAB, a popular software platform for numerical computation. As illustrated via several tests, the randomized algorithms for low-rank approximation outperform or at least match the classical deterministic techniques (such as Lanczos iterations run to convergence) in basically all respects: accuracy, computational efficiency (both speed and memory usage), ease-of-use, parallelizability, and reliability. However, the classical procedures remain the methods of choice for estimating spectral norms and are far superior for calculating the least singular values and corresponding singular vectors (or singular subspaces).Recent years have witnessed intense development of randomized methods for low-rank approximation. These methods target principal component analysis and the calculation of truncated singular value decompositions. The present article presents an essentially black-box, foolproof implementation for Mathworks' MATLAB, a popular software platform for numerical computation. As illustrated via several tests, the randomized algorithms for low-rank approximation outperform or at least match the classical deterministic techniques (such as Lanczos iterations run to convergence) in basically all respects: accuracy, computational efficiency (both speed and memory usage), ease-of-use, parallelizability, and reliability. However, the classical procedures remain the methods of choice for estimating spectral norms and are far superior for calculating the least singular values and corresponding singular vectors (or singular subspaces). Recent years have witnessed intense development of randomized methods for low-rank approximation. These methods target principal component analysis and the calculation of truncated singular value decompositions. The present article presents an essentially black-box, foolproof implementation for Mathworks' MATLAB, a popular software platform for numerical computation. As illustrated via several tests, the randomized algorithms for low-rank approximation outperform or at least match the classical deterministic techniques (such as Lanczos iterations run to convergence) in basically all respects: accuracy, computational efficiency (both speed and memory usage), ease-of-use, parallelizability, and reliability. However, the classical procedures remain the methods of choice for estimating spectral norms and are far superior for calculating the least singular values and corresponding singular vectors (or singular subspaces). |
| Author | Stanton, Kelly P Tygert, Mark Li, Huamin Linderman, George C Kluger, Yuval Szlam, Arthur |
| Author_xml | – sequence: 1 givenname: Huamin surname: Li fullname: Li, Huamin organization: Program in Applied Mathematics, 51 Prospect St., Yale University, New Haven, CT 06510 – sequence: 2 givenname: George C surname: Linderman fullname: Linderman, George C organization: Program in Applied Mathematics, 51 Prospect St., Yale University, New Haven, CT 06510 – sequence: 3 givenname: Arthur surname: Szlam fullname: Szlam, Arthur organization: Facebook, 8th floor, 770 Broadway, New York, NY 10003 – sequence: 4 givenname: Kelly P surname: Stanton fullname: Stanton, Kelly P organization: Yale University, School of Medicine, Department of Pathology, Suite 505L, 300 George St., New Haven, CT 06520 – sequence: 5 givenname: Yuval surname: Kluger fullname: Kluger, Yuval organization: Yale University, School of Medicine, Department of Pathology, Suite 505L, 300 George St., New Haven, CT 06520 – sequence: 6 givenname: Mark surname: Tygert fullname: Tygert, Mark organization: Facebook, 1 Facebook Way, Menlo Park, CA 94025 |
| BackLink | https://www.ncbi.nlm.nih.gov/pubmed/28983138$$D View this record in MEDLINE/PubMed |
| BookMark | eNpFkEtLAzEYRbOo2FbFfyBZuhn9kpl0Enel-CgUFNGViyFPjeQxTqaL-ustWHF1N-ceLneOJikni9A5gStCGnZdAzTA6gmaAQhe1QxgiualfAIAJS05RlPKBa9JzWfobRne8-DHj4hFS27wMuF17IONNo1y9Dnh7LDEzzKZHP23Nfi_4PKAnwaftO9lwKsc-_2QNO4dMuyKL6foyMlQ7NkhT9Dr3e3L6qHaPN6vV8tNpeuFGCsuSWMUIVoLUHqhwGjXaKY4papttdGMKO6gdYw3giptwFptW6sdI0ZZR0_Q5a-3H_LX1paxi75oG4JMNm9LR0TDW8YYXezRiwO6VdGarh98lMOu-zuE_gC7bWGt |
| CitedBy_id | crossref_primary_10_1002_gepi_22136 crossref_primary_10_1088_2632_2153_ab8240 crossref_primary_10_1002_2017WR021870 crossref_primary_10_1007_s10915_023_02411_2 crossref_primary_10_1515_mcma_2023_2012 crossref_primary_10_1137_23M1558537 crossref_primary_10_1137_22M1538648 crossref_primary_10_1155_2021_9840335 crossref_primary_10_1017_jpr_2020_21 crossref_primary_10_1109_TMTT_2021_3090798 crossref_primary_10_1145_3506691 crossref_primary_10_1137_17M1111590 crossref_primary_10_1021_jasms_4c00314 crossref_primary_10_1137_17M1141977 crossref_primary_10_1137_22M154079X crossref_primary_10_1017_S0962492920000021 crossref_primary_10_1137_23M1548323 crossref_primary_10_1109_TPAMI_2018_2839198 crossref_primary_10_1137_18M118966X crossref_primary_10_1007_s11227_022_04524_1 crossref_primary_10_1007_s13253_020_00415_1 crossref_primary_10_1137_20M1355720 crossref_primary_10_1137_21M1466244 crossref_primary_10_1137_17M1138480 crossref_primary_10_3390_a15060190 crossref_primary_10_1007_s10618_023_00976_y crossref_primary_10_1038_s41592_018_0308_4 crossref_primary_10_1145_3660629 crossref_primary_10_1002_nme_6068 crossref_primary_10_1186_s13059_019_1900_3 crossref_primary_10_1186_s13059_019_1639_x |
| ContentType | Journal Article |
| DBID | NPM 7X8 |
| DOI | 10.1145/3004053 |
| DatabaseName | PubMed MEDLINE - Academic |
| DatabaseTitle | PubMed MEDLINE - Academic |
| DatabaseTitleList | MEDLINE - Academic PubMed |
| Database_xml | – sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database |
| DeliveryMethod | no_fulltext_linktorsrc |
| Discipline | Sciences (General) Computer Science |
| ExternalDocumentID | 28983138 |
| Genre | Journal Article |
| GrantInformation_xml | – fundername: NCI NIH HHS grantid: R01 CA158167 – fundername: NHGRI NIH HHS grantid: F30 HG010102 – fundername: NIGMS NIH HHS grantid: T32 GM007205 – fundername: NHGRI NIH HHS grantid: R01 HG008383 |
| GroupedDBID | --Z -DZ -~X .DC 23M 2FS 4.4 5GY 5VS 6J9 6OB 85S 8US 9M8 AAIKC AAKMM AALFJ AAMNW AAYFX AAYOK ABFSI ABPPZ ABTAH ACGFO ACGOD ACIWK ACM ACNCT ADBCU ADL ADPZR AEBYY AENEX AENSD AFFNX AFWIH AFWXC AGHSJ AI. AIKLT ALMA_UNASSIGNED_HOLDINGS ASPBG AVWKF BDXCO CCLIF CS3 D0L E.L EBS EJD FEDTE GUFHI HF~ HGAVV H~9 I07 IAO ICD IEA IGS IOF ITC LHSKQ MS~ MVM NHB NPM OHT P1C P2P PKN PQQKQ RNS ROL RXW TAE TWZ U5U UAO UHB UPT VH1 W7O X6Y XJT XOL ZCA ZY4 7X8 AEFXT AEJOY AETEA AKRVB AMVHM |
| ID | FETCH-LOGICAL-c369t-8a14db11cc90bc6b0dcf4c5b822b77cdc51b8f07f58492bcd0eece7ecf51dbef2 |
| IEDL.DBID | 7X8 |
| ISICitedReferencesCount | 21 |
| ISICitedReferencesURI | http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000395512100011&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| ISSN | 0098-3500 |
| IngestDate | Fri Sep 05 09:28:07 EDT 2025 Wed Feb 19 02:44:18 EST 2025 |
| IsPeerReviewed | true |
| IsScholarly | true |
| Issue | 3 |
| Keywords | SVD Algorithms Performance singular value decomposition PCA Principal component analysis |
| Language | English |
| LinkModel | DirectLink |
| MergedId | FETCHMERGED-LOGICAL-c369t-8a14db11cc90bc6b0dcf4c5b822b77cdc51b8f07f58492bcd0eece7ecf51dbef2 |
| Notes | ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 23 |
| PMID | 28983138 |
| PQID | 1948755526 |
| PQPubID | 23479 |
| ParticipantIDs | proquest_miscellaneous_1948755526 pubmed_primary_28983138 |
| PublicationCentury | 2000 |
| PublicationDate | 2017-01-01 |
| PublicationDateYYYYMMDD | 2017-01-01 |
| PublicationDate_xml | – month: 01 year: 2017 text: 2017-01-01 day: 01 |
| PublicationDecade | 2010 |
| PublicationPlace | United States |
| PublicationPlace_xml | – name: United States |
| PublicationTitle | ACM transactions on mathematical software |
| PublicationTitleAlternate | ACM Trans Math Softw |
| PublicationYear | 2017 |
| SSID | ssj0002171 |
| Score | 2.4499264 |
| Snippet | Recent years have witnessed intense development of randomized methods for low-rank approximation. These methods target principal component analysis and the... |
| SourceID | proquest pubmed |
| SourceType | Aggregation Database Index Database |
| Title | Algorithm 971: An Implementation of a Randomized Algorithm for Principal Component Analysis |
| URI | https://www.ncbi.nlm.nih.gov/pubmed/28983138 https://www.proquest.com/docview/1948755526 |
| Volume | 43 |
| WOSCitedRecordID | wos000395512100011&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D |
| hasFullText | |
| inHoldings | 1 |
| isFullTextHit | |
| isPrint | |
| link | http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1LS8QwEA6-Dl7U9f0kggc9FJO26cOLLOLixWURhQUPSzJJdMFtV7t68Nc7aVP2JAheSi_TKZlJ5ptJ8g0hZ0nMpQS0QBLJPED8rwMER3kABsODFIlKIK6bTaT9fjYc5gNfcKv8scp2TawXal2Cq5FfYrKN0FqIMLmevgeua5TbXfUtNBbJcoRQxh3pSodztnCE27zlzIwEY82lWcwAxKUjmmKuI_JvuLKOL731__7ZBlnzyJJ2G1fokAVTbJL1tmsD9ZN4k3T8W0XPPef0xRZ57r694DdnrxOap_yKdgta8wZP_NWkgpaWSvogC11Oxt9G07kAwl46aGr2qN8pLAsUoy3fyTZ56t0-3twFvu9CAFGSz4JM8lgrzgFypiBRTIONQSjEEipNQYPgKrMstQhe8lCBZsaASQ1YwbUyNtwhSwVq2iPUhiAjG2rgTMaZ0ZJpIxHBCIOwwPJ0n5y2YzpCv3abFbIw5Wc1mo_qPtltDDOaNgQcI0wSM7R3dvAH6UOyGrpIXFdNjsiyxVltjskKfM3G1cdJ7TD47A_ufwAxLMye |
| linkProvider | ProQuest |
| openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Algorithm+971%3A+An+Implementation+of+a+Randomized+Algorithm+for+Principal+Component+Analysis&rft.jtitle=ACM+transactions+on+mathematical+software&rft.au=Li%2C+Huamin&rft.au=Linderman%2C+George+C&rft.au=Szlam%2C+Arthur&rft.au=Stanton%2C+Kelly+P&rft.date=2017-01-01&rft.issn=0098-3500&rft.volume=43&rft.issue=3&rft_id=info:doi/10.1145%2F3004053&rft.externalDBID=NO_FULL_TEXT |
| thumbnail_l | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0098-3500&client=summon |
| thumbnail_m | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0098-3500&client=summon |
| thumbnail_s | http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0098-3500&client=summon |