Tesi etd-03122024-153814 |
Link copiato negli appunti
Tipo di tesi
Tesi di laurea magistrale
Autore
SIMONI, WILLIAM
URN
etd-03122024-153814
Titolo
Graph Learning for Network Quantification
Dipartimento
INFORMATICA
Corso di studi
INFORMATICA
Relatori
relatore Prof. Micheli, Alessio
relatore Dott. Sebastiani, Fabrizio
relatore Dott. Podda, Marco
relatore Dott. Sebastiani, Fabrizio
relatore Dott. Podda, Marco
Parole chiave
- Network Quantification
- Graph Learning
Data inizio appello
12/04/2024
Consultabilità
Non consultabile
Data di rilascio
12/04/2094
Riassunto
Among the various challenges that machine learning addresses, this thesis focuses on the problem of graph quantification. Quantification is the task of estimating the class prevalence values (or
class priors) within a sample of unlabelled data points that exhibit Dataset Shift. Graph (or Network) quantification extends this task to the graph setting, which is characterized by data points that are interconnected with each other.
The thesis analyzes the impact of quantification methods and graph learning methods in the context of graph quantification. Moreover, it provides a new state-of-the-art method, GESN-SLD, that outperforms all the previously presented methods in the literature.
class priors) within a sample of unlabelled data points that exhibit Dataset Shift. Graph (or Network) quantification extends this task to the graph setting, which is characterized by data points that are interconnected with each other.
The thesis analyzes the impact of quantification methods and graph learning methods in the context of graph quantification. Moreover, it provides a new state-of-the-art method, GESN-SLD, that outperforms all the previously presented methods in the literature.
File
| Nome file | Dimensione |
|---|---|
Tesi non consultabile. |
|