University of Patras
Μεθοδολογικό πλαίσιο υποστήριξης της εξόρυξης γνώσης από δεδομένα με την χρήση αρχών της πολυκριτήριας ανάλυσης αποφάσεων
Abstract
dc:descriptionData mining is a new and advancing technology that helps corporations to focus on the most important pieces of information stored in their data warehouses. In particular, data mining searches for hidden patterns and is able to discover information that otherwise could be missed or overlooked by experts. During the last years, a series of data mining algorithms has been developed. These algorithms are based on different methodological patterns and they can be implemented to solve a large variety of applications. However, the effort to build more advanced and efficient data mining algorithms has never stopped. The goal of this PhD thesis is to significantly contribute to the above effort by enhancing and improving the theoretical framework of existing data mining algorithms. More specifically, a different theoretical perspective is introduced, that is based on concepts and procedures of multicriteria analysis and in particular the ELECTRE I method of the outranking relations theory. Consequently, the PhD thesis develops a new methodological framework for data mining that can be incorporated to existing algorithms. This incorporation essentially develops new, more effective and accurate data mining algorithms, for a series of tasks and applications. In particular, the proposed methodological framework was applied, properly modified, to the tasks of classification and clustering, through the CLEDM and CLEKMODES methods, respectively. The good results of these methods in a series of widely used databases, and the perspective of expanding the new methodological framework to other data mining tasks as well, are able to introduce a new “hybrid” research field. This research field has the potential of producing better data mining algorithms and furthermore the potential to thoroughly explore and further formalize the interaction of data mining and multicriteria analysis.
Degree
thesis:*- Grantor dc:publisher
- University of Patras
- Year dc:date
- 2009
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Mastrogiannis, Nikolaos
- Μαστρογιάννης, Νικόλαος
Subjects
dc:subject × 18- Εξόρυξη δεδομένων
- Βάσεις δεδομένων
- Πολυκριτήρια ανάλυση αποφάσεων
- Μέθοδος ELECTRE I
- Αλγόριθμοι
- Ταξινόμηση
- Ομαδοποίηση
- Data mining
- Databases
- Multicriteria analysis
- ELECTRE I method
- Algorithms
- Classification
- Clustering
- Φυσικές Επιστήμες
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Natural Sciences
- Computer and Information Sciences
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/32758