Gradient-based methods for solving non-negative matrix tri-factorization problem : Code and data

Povh, Janez (author), Hribar, Rok (author)

Code and data used in paper "Gradient-based methods for solving non-negative matrix tri-factorization problem". GitLab: symmetric NMTF, Project ID: 1360, 3 branches, total 5.4 GB files. History: 06 Jul, 2020 3 commits. ZIP in this repository contains code and small amount of data used in code examples. Full data is available from GitLab.

• Language: English
• Keywords: non-negative matrix factorization, p-matching problem, data co-clustering, patients sub-typing, drug repurposing, high-performance computing
• Organization: FS - Faculty of Mechanical Engineering
• Year: 2020
• PID: 20.500.12556/RUL-117359

Copyright

No license. All rights reserved.

Note

With regard to the type and amount of data one requires, three options are available: 1. code and small amount of data used in code examples (~1M): git clone --single-branch https://repo.ijs.si/hribarr/symmetric-nmtf.git; 2. code and real-world data used in the paper (~210M): git clone --single-branch https://repo.ijs.si/hribarr/symmetric-nmtf.git -b real-data; 3. code, real-world data and synthetic data used in the paper (~14G): git clone --single-branch https://repo.ijs.si/hribarr/symmetric-nmtf.git -b full-data.

• Publication date in RUL: 08.07.2020

Secondary language

• Language: Slovenian
• Keywords: nenegativna matrična faktorizacija, problem p-prirejanja, so-gručenje podatkov, serotipizacija pacientov, nova uporabnost zdravil, super-računalništvo

Projects

• Funder: ARRS - Slovenian Research Agency
• Project number: J1-8155
• Name: ZLIVANJE BIOMEDICINSKIH PODATKOV Z UPORABO NENEGATIVNE MATRIČNE TRI-FAKTORIZACIJE