ADMM algorithms for matrix completion problem in noisy settings
Matrix completion (MC) is a fundamental linear algebra problem to fully recover a low-rank matrix from its incomplete data. It is widely applied in machine learning and statistics, varied from wireless communication, image compression to collaborative filtering. Meanwhile, Alternating Direction Meth...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2021
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/148502 |
| Summary: | Matrix completion (MC) is a fundamental linear algebra problem to fully recover a low-rank matrix from its incomplete data. It is widely applied in machine learning and statistics, varied from wireless communication, image compression to collaborative filtering. Meanwhile, Alternating Direction Method of Multiplier is a straightforward but effective algorithm for distributed convex optimization. In this work, we will study ADMM in application to matrix completion problem in the noisy setting. Two modified algorithms for noisy matrix completion problem are proposed. Convergence results of these algorithms will be discussed and numerical experiments are conducted to examine the performance of the new algorithms. |
|---|