Abstract
Correlation matrices play a crucial role in assessing credit loan eligibility, where missing data can significantly impact the results. In this paper, we address these challenges by employing different approaches that tackle two aspects of the problem: transforming a partially specified matrix into a fully specified correlation matrix and completing a partially specified matrix while preserving the known data. The correlation problem is effectively solved using the interior point primal-dual path-following method. To provide comprehensive insights, this paper presents a thorough comparative analysis of our proposed methods, comparing them with approaches that utilize the modified alternating projection method. The efficacy and computational efficiency of the proposed methods are evaluated by analyzing numerical results.
| Original language | English |
|---|---|
| Pages (from-to) | 1007-1027 |
| Number of pages | 21 |
| Journal | Numerical Algebra, Control and Optimization |
| Volume | 15 |
| Issue number | 4 |
| DOIs | |
| State | Published - Dec 2025 |
Bibliographical note
Publisher Copyright:© 2025, American Institute of Mathematical Sciences. All rights reserved.
Keywords
- Correlation matrix
- alternating projection method
- credit loan eligibility
- matrix approximation
- matrix completions
- semidefinite programming
ASJC Scopus subject areas
- Algebra and Number Theory
- Control and Optimization
- Applied Mathematics
Fingerprint
Dive into the research topics of 'ON CORRELATION MATRIX COMPLETION PROBLEM WITH APPLICATION TO CREDIT LOAN ELIGIBILITY'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver