University of the Aegean
Ταυτόχρονη αναγνώριση τάξης και παραμέτρων μοντέλων χρονοσειρών με χρήση πολυμοντελικού διαμελισμού
Abstract
dc:descriptionThis thesis deals with the extension of the multi-model partitioning theory to multivariate models and its application to real cases such as computer networks, electric load demand forecasting and the variation of the grounding resistance during the year. In the first chapter the multi-model partitioning theory is used for simultaneous order and parameter estimation of multivariate autoregressive models as an extension to the one proposed for the scalar case in. It is also shown that the method is also successful in tracking model order changes, in real time [57]. The second chapter deals with adaptive network anomaly detection using bandwidth utilization data. The proposed method has two advantages: it is based on a powerful multi-model partitioning filter, (MMPF) proposed by Lainiotis [33-35], known for its stability and well established in identification and modeling; secondly, it uses easy to find and collect datasets. Bandwidth use is the most common set of data and almost all network administrators monitor the bandwidth utilization for their servers, LAN/VLAN users, and network connections. The applied adaptive multi-model partitioning filter identified successfully the unusual activities. The method can perform equally well in real-time using the sampling interval required by the network monitoring programs. Chapter three applies the multi-model partitioning filter in order to perform simultaneous order and parameter estimation of Multivariate (MV) ARMA (AutoRegressive Moving Average) models under the presence of noise. The proposed method successfully selects the correct model order in very few steps and identifies very accurately the ARMA parameters. Comparison with other established order selection criteria shows that the method needs the shortest data set for successful order identification and accurate parameter estimation for all the simulated models, whereas the other criteria require longer data sets as the model order increases. Finally, the method is capable of tracking, in real time, any model order changes. Chapter 4 deals with efficient modeling and forecasting of the electricity demand load using real data. Load anomalies such as unexpected peaks that may appear during the summer or unexpected faults (blackouts) are also modeled. If the load pattern does not match the normal behavior of the load, an anomaly is detected furthermore, when the pattern matches a known case of anomaly, the type of anomaly is identified. The applied adaptive multi-model partitioning filter identified successfully both normal periodic behavior and any unusual activity of the electric grid. This work can be useful in the studies that concern electricity consumption and electricity prices forecasts giving the possibility to the electricity providers, retailers and regulatory authorities to supply uninterrupted energy at a low cost. Chapter 5 addresses the problem of modeling the variation of the grounding resistance during the year. For validation purposes the proposed method was compared with three other established order selection criteria (AICC [25-27], AIC [8], BIC [19]) and presented very good results. The proposed method can be extremely useful in the design of electrical installations, since the variation of the grounding resistance during the year affects significantly power systems performance and must be definitely considered.
Degree
thesis:*- Grantor dc:publisher
- University of the Aegean
- Year dc:date
- 2008
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Pappas, Stylianos
- Παππάς, Στυλιανός
Subjects
dc:subject × 24- Αλγόριθμοι πολυμοντελικού διαμελισμού
- Πολυμεταβλητά ακολουθιακά μοντέλα
- Αναγνώριση τάξης
- Αναγνώριση παραμέτρων
- Μοντέλα μεικτών διεργασιών
- Νευρωνικά δίκτυα
- Εκτίμηση ηλεκτρικού φορτίου
- Εκτίμηση φορτίου δικτύου υπολογιστών
- Εκτίμηση μεταβολής της αντίστασης του εδάφους
- Φίλτρα Kalman
- Multimodel partitioning algorithms
- Multivariate auto-regressive (AR) models
- Order identification
- Parameter estimation
- Multivariate autoregressive moving average (ARMA) models
- Neural networks
- Electric load estimation
- Computer network anomaly detection
- Prediction of the grounding resistance change
- Kalman filters
- Επιστήμες Μηχανικού και Τεχνολογία
- Επιστήμη Ηλεκτρολόγου Μηχανικού, Ηλεκτρονικού Μηχανικού, Μηχανικού Η/Υ
- Engineering and Technology
- Electrical Engineering, Electronic Engineering, Information Engineering
Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/26056