Tesi etd-05222024-153144
Link copiato negli appunti
Type of thesis
Corso Ordinario Secondo Livello
Author
MICHELIS, FILIPPO
URN
etd-05222024-153144
Title
Nonnegative Matrix Factorization with Sparse Inverse Covariance
Structure
Classe Scienze Sociali
Course
SCIENZE ECONOMICHE E MANAGERIALI - SCIENZE ECONOMICHE E MANAGERIALI
Committee
relatore Prof.ssa CHIAROMONTE, FRANCESCA
Presidente Prof. IRALDO, FABIO
Tutor Prof. BELLINI, NICOLA
Membro Prof. FAGIOLO, GIORGIO
Membro Prof. TESTA, FRANCESCO
Membro Prof.ssa ANNUNZIATA, ELEONORA
Presidente Prof. IRALDO, FABIO
Tutor Prof. BELLINI, NICOLA
Membro Prof. FAGIOLO, GIORGIO
Membro Prof. TESTA, FRANCESCO
Membro Prof.ssa ANNUNZIATA, ELEONORA
Keywords
- Nessuna parola chiave trovata
Exam session start date
14/06/2024;
Availability
completa
Abstract
Nonnegative matrix factorization (NMF) has become increasingly popular as a powerful method for clustering and latent factor modeling. It decomposes data into a combination of latent fundamental parts, making it effective for dimensionality reduction. Previous research has primarily focused on obtaining meaningful parts, by manifold regularization, sparseness constraints, or orthogonality, without ever modeling co-occurrence patterns directly. A new NMF method is presented that achieves meaningful parts considering how they co-occur. Attention is shifted to the relationships within parts by modeling the conditional independence relationships among them. To accomplish this, a novel regularization term is proposed that induces sparsity on the inverse covariance matrix of the latent factors. The method performance is assessed through extensive simulations and a real data case study.
Files
| Nome file | Dimensione |
|---|---|
| Thesis_N...iance.pdf | 2.20 Mb |
Contatta l'autore |
|