Aristotle University Of Thessaloniki (AUTH)
Τεχνικές μηχανικής μάθησης για πρόβλεψη πολλαπλών μεταβλητών
Abstract
dc:descriptionThis thesis studies a category of machine learning problems where the aim is to construct models that predict multiple target variables based on a common set of input variables. We developed a complete software for learning from multi-label data and showed that it is useful as both a scientific as well as a practical tool. We also studied the problem of multi-target regression and proposed novel techniques that treat target variables as additional input variables. The proposed methods manage to successfully model dependencies in the output space, thus attaining significantly better accuracy than both the baseline approach as well as four state-of-the-art methods from the literature. Finally, the thesis focused at a popular multi-label classification problem, automated image annotation. We proposed new vectorized image representation methods that attain state-of-the-art accuracy while also being highly scalable.
Degree
thesis:*- Grantor dc:publisher
- Aristotle University Of Thessaloniki (AUTH)
- Year dc:date
- 2016
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Spyromitros-Xioufis, Eleftherios
- Σπυρομήτρος-Ξιούφης, Ελευθέριος
Subjects
dc:subject × 16- Ταξινόμηση πολλαπλών ετικετών
- Παλινδρόμηση πολλαπλών στόχων
- Ταξινόμηση εικόνων
- Μηχανική μάθηση
- Multi-label classification
- Multi-target regression
- Image classification
- Machine learning
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Natural Sciences
- Computer and Information Sciences
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
dc:rights- Statement dc:rights
-
- BY_SA
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/38512