Thesis (Ph.D.)--McMaster University, 2006. / Supervisor: Susan J. Elliott. Includes bibliographical references (leaves 166-171).

This dissertation addresses the problem of distributed anomaly detection in Wireless Sensor Networks (WSN). A challenge of designing such systems is that the sensor nodes are battery powered, often have different capabilities and generally operate in dynamic environments. Programming such sensor nodes at a large scale can be a tedious job if the system is not carefully designed. Data modeling in distributed systems is important for determining the normal operation mode of the system. Being able to model the expected sensor signatures for typical operations greatly simplifies the human designer’s job by enabling the system to autonomously characterize the expected sensor data streams. This, in turn, allows the system to perform autonomous anomaly detection to recognize when unexpected sensor signals are detected. This type of distributed sensor modeling can be used in a wide variety of sensor networks, such as detecting the presence of intruders, detecting sensor failures, and so forth. The advantage of this approach is that the human designer does not have to characterize the anomalous signatures in advance. The contributions of this approach include: (1) providing a way for a WSN to autonomously model sensor data with no prior knowledge of the environment; (2) enabling a distributed system to detect anomalies in both sensor signals and temporal events online; (3) providing a way to automatically extract semantic labels from temporal sequences; (4) providing a way for WSNs to save communication power by transmitting compressed temporal sequences; (5) enabling the system to detect time-related anomalies without prior knowledge of abnormal events; and, (6) providing a novel missing data estimation method that utilizes temporal and spatial information to replace missing values. The algorithms have been designed, developed, evaluated, and validated experimentally in synthesized data, and in real-world sensor network applications.

The separation of natural and anthropogenically caused climatic changes is an important task of contemporary climate research. For this purpose, a detailed knowledge of the natural variability of the climate during warm stages is a necessary prerequisite. Beside model simulations and historical documents, this knowledge is mostly derived from analyses of so-called climatic proxy data like tree rings or sediment as well as ice cores. In order to be able to appropriately interpret such sources of palaeoclimatic information, suitable approaches of statistical modelling as well as methods of time series analysis are necessary, which are applicable to short, noisy, and non-stationary uni- and multivariate data sets. Correlations between different climatic proxy data within one or more climatological archives contain significant information about the climatic change on longer time scales. Based on an appropriate statistical decomposition of such multivariate time series, one may estimate dimensions in terms of the number of significant, linear independent components of the considered data set. In the presented work, a corresponding approach is introduced, critically discussed, and extended with respect to the analysis of palaeoclimatic time series. Temporal variations of the resulting measures allow to derive information about climatic changes. For an example of trace element abundances and grain-size distributions obtained near the Cape Roberts (Eastern Antarctica), it is shown that the variability of the dimensions of the investigated data sets clearly correlates with the Oligocene/Miocene transition about 24 million years before present as well as regional deglaciation events. Grain-size distributions in sediments give information about the predominance of different transportation as well as deposition mechanisms. Finite mixture models may be used to approximate the corresponding distribution functions appropriately. In order to give a complete description of the statistical uncertainty of the parameter estimates in such models, the concept of asymptotic uncertainty distributions is introduced. The relationship with the mutual component overlap as well as with the information missing due to grouping and truncation of the measured data is discussed for a particular geological example. An analysis of a sequence of grain-size distributions obtained in Lake Baikal reveals that there are certain problems accompanying the application of finite mixture models, which cause an extended climatological interpretation of the results to fail. As an appropriate alternative, a linear principal component analysis is used to decompose the data set into suitable fractions whose temporal variability correlates well with the variations of the average solar insolation on millenial to multi-millenial time scales. The abundance of coarse-grained material is obviously related to the annual snow cover, whereas a significant fraction of fine-grained sediments is likely transported from the Taklamakan desert via dust storms in the spring season. / Die Separation natürlicher und anthropogen verursachter Klimaänderungen ist eine bedeutende Aufgabe der heutigen Klimaforschung. Hierzu ist eine detaillierte Kenntnis der natürlichen Klimavariabilität während Warmzeiten unerlässlich. Neben Modellsimulationen und historischen Aufzeichnungen spielt hierfür die Analyse von sogenannten Klima-Stellvertreterdaten eine besondere Rolle, die anhand von Archiven wie Baumringen oder Sediment- und Eisbohrkernen erhoben werden. Um solche Quellen paläoklimatischer Informationen vernünftig interpretieren zu können, werden geeignete statistische Modellierungsansätze sowie Methoden der Zeitreihenanalyse benötigt, die insbesondere auf kurze, verrauschte und instationäre uni- und multivariate Datensätze anwendbar sind. Korrelationen zwischen verschiedenen Stellvertreterdaten eines oder mehrerer klimatologischer Archive enthalten wesentliche Informationen über den Klimawandel auf großen Zeitskalen. Auf der Basis einer geeigneten Zerlegung solcher multivariater Zeitreihen lassen sich Dimensionen schätzen als die Zahl der signifikanten, linear unabhängigen Komponenten des Datensatzes. Ein entsprechender Ansatz wird in der vorliegenden Arbeit vorgestellt, kritisch diskutiert und im Hinblick auf die Analyse von paläoklimatischen Zeitreihen weiterentwickelt. Zeitliche Variationen der entsprechenden Maße erlauben Rückschlüsse auf klimatische Veränderungen. Am Beispiel von Elementhäufigkeiten und Korngrößenverteilungen des Cape-Roberts-Gebietes in der Ostantarktis wird gezeigt, dass die Variabilität der Dimension der untersuchten Datensätze klar mit dem Übergang vom Oligozän zum Miozän vor etwa 24 Millionen Jahren sowie regionalen Abschmelzereignissen korreliert. Korngrößenverteilungen in Sedimenten erlauben Rückschlüsse auf die Dominanz verschiedenen Transport- und Ablagerungsmechanismen. Mit Hilfe von Finite-Mixture-Modellen lassen sich gemessene Verteilungsfunktionen geeignet approximieren. Um die statistische Unsicherheit der Parameterschätzung in solchen Modellen umfassend zu beschreiben, wird das Konzept der asymptotischen Unsicherheitsverteilungen eingeführt. Der Zusammenhang mit dem Überlapp der einzelnen Komponenten und aufgrund des Abschneidens und Binnens der gemessenen Daten verloren gehenden Informationen wird anhand eines geologischen Beispiels diskutiert. Die Analyse einer Sequenz von Korngrößenverteilungen aus dem Baikalsee zeigt, dass bei der Anwendung von Finite-Mixture-Modellen bestimmte Probleme auftreten, die eine umfassende klimatische Interpretation der Ergebnisse verhindern. Stattdessen wird eine lineare Hauptkomponentenanalyse verwendet, um den Datensatz in geeignete Fraktionen zu zerlegen, deren zeitliche Variabilität stark mit den Schwankungen der mittleren Sonneneinstrahlung auf der Zeitskala von Jahrtausenden bis Jahrzehntausenden korreliert. Die Häufigkeit von grobkörnigem Material hängt offenbar mit der jährlichen Schneebedeckung zusammen, während feinkörniges Material möglicherweise zu einem bestimmten Anteil durch Frühjahrsstürme aus der Taklamakan-Wüste herantransportiert wird.

Einleitung & Problemstellung: Beschwerden nach Beschleunigungsverletzungen der Halswirbelsäule sind oft nur unzureichend einzuordnen und diagnostizierbar. Eine eindeutige Diagnostik ist jedoch für eine entsprechende Therapie wie auch möglicherweise entstehende versicherungsrechtliche Forderungen notwendig. Die Entwicklung eines geeigneten Diagnoseverfahrens liegt damit im Interesse von Betroffenen wie auch Kostenträgern. Neben Störungen der Weichteilgewebe ist fast immer die Funktion der Halsmuskulatur in Folge eines Traumas beeinträchtigt. Dabei wird vor allem die sensorische Funktion der HWS-Muskulatur, die an der Regulation des Gleichgewichts beteiligt ist, gestört. In Folge dessen kann angenommen werden, dass es zu einer Beeinträchtigung der Gleichgewichtsregulation kommt. Die Zielstellung der Arbeit lautete deshalb, die möglicherweise gestörte Gleichgewichtsregulation nach einem Trauma im HWS-Bereich apparativ zu erfassen, um so die Verletzung eindeutig diagnostizieren zu können. Methodik: Unter Verwendung eines posturographischen Messsystems mit Kraftmomentensensorik wurden bei 478 Probanden einer Vergleichsgruppe und bei 85 Probanden eines Patientenpools Kraftmomente unter der Fußsohle als Äußerung der posturalen Balanceregulation aufgezeichnet. Die gemessenen Balancezeitreihen wurden nichtlinear analysiert, um die hohe Variabilität der Gleichgewichtsregulation optimal zu beschreiben. Über die dabei gewonnenen Parameter kann überprüft werden, ob sich spezifische Unterschiede im Schwankungsverhalten anhand der plantaren Druckverteilung zwischen HWS-Traumatisierten und den Probanden der Kontrollgruppe klassifizieren lassen. Ergebnisse: Die beste Klassifizierung konnte dabei über Parameter erzielt werden, die das Schwankungsverhalten in Phasen beschreiben, in denen die Amplitudenschwankungen relativ gering ausgeprägt waren. Die Analysen ergaben signifikante Unterschiede im Balanceverhalten zwischen der Gruppe HWS-traumatisierter Probanden und der Vergleichsgruppe. Die höchsten Trennbarkeitsraten wurden dabei durch Messungen im ruhigen beidbeinigen Stand mit geschlossenen Augen erzielt. Diskussion: Das posturale Balanceverhalten wies jedoch in allen Messpositionen eine hohe individuelle Varianz auf, so dass kein allgemeingültiges Schwankungsmuster für eine Gruppengesamtheit klassifiziert werden konnte. Eine individuelle Vorhersage der Gruppenzugehörigkeit ist damit nicht möglich. Die verwendete Messtechnik und die angewandten Auswerteverfahren tragen somit zwar zu einem Erkenntnisgewinn und zur Beschreibung des Gleichgewichtsverhaltens nach HWS-Traumatisierung bei. Sie können jedoch zum derzeitigen Stand für den Einzelfall keinen Beitrag zu einer eindeutigen Bestimmung eines Schleudertraumas leisten. / Introduction & Problem definition: Disorders after acceleration injuries of the cervical spine can often be classified and diagnosed only inadequately. But an explicit diagnosis is necessary as a basis for an adequate therapy as well as for possibly arising demands pursuant to insurance law. The development of suitable diagnosis methods is in the interest of patients as well as the cost units. Apart from disorders of the soft tissues there are almost always impairments of the function of the neck musculature. Particularly the sensory function of the cervical spine musculature, which participates in the regulation of the equilibrium, is disturbed by that. As a result in can be assumed that the postural control is also disturbed. Therefore the aim of this study was to examine the possibly disturbed postural motor balance after a whiplash injury of the cervical spine with the help of apparatus-supported methods to be able to unambigiously diagnose. Methods: postural measuring system based on the force-moment sensortechnique was used to record the postural balance regulation of 478 test persons and 85 patients which had suffered a whiplash injury. Data analysis was accomplished by linear as well as by nonlinear time series methods in order to characterise the balance regulation in an optimal way. Thus it can be determined whether there can be classified specific differences in the plantar pressure distribution covering patients with a whiplash injury and the test persons of the control group. Results: The best classification could be achieved by parameters which describe the variation of the postural balance regulation in phases in which the differences of the amplitudes of the plantar pressure distribution were relatively small. The analyses showed significant differences in the postural motor balance between the group of patients with whiplash injuries and the control group. The most significant differences (highest discriminate rates) could be observed by measurements in both-legged position with closed eyes. Discussion: Although the results achieved support the hypothesis mentioned above, is must be conceded that the postural motor balance showed a high individual variation in all positions of measurement. Therefore no universal variation model could be classified for the entirety of either group. This way an individual forecast of the group membership is impossible. As a result the measurement technology being used and the nonlinear time series analyses can contribute to the gain of knowledge and to the description of the regulation of postural control after whiplash injury. But at present they cannot contribute to an explicit determination of a whiplash injury for a particular case.

The Greenland Ice Sheet is the largest ice sheet in the northern hemisphere. Ongoing melting of the ice sheet, resulting in increased mass loss relative to the longer term trend, has raised concerns about the stability of the ice sheet. Melt water generated at the surface is temporarily stored in supra-glacial lakes on the ice sheet. Connections between melt water generation, storage and ice sheet dynamics highlight the importance of the surface hydrological system. In this thesis different methods are used that improve our ability to observe the supra-glacial lake system on the west Greenland Ice Sheet. This region of the Greenland Ice Sheet has the most extensive supra-glacial hydrological system with a dense network of streams connecting lakes that can exceed several square kilometres in area. Synthetic Aperture Radar (SAR) and visible-near infrared (VNIR) images are used to explore the potential of different sensor systems for regular observations of the supra-glacial lakes. SAR imagery is found to be a useful complement to VNIR data. VNIR data from moderate resolution sensors are preferred as these provide high temporal resolution data, ameliorating problems with cloud cover. The dynamic nature of the lakes makes automated classification difficult and manual mapping has been widely used. Here a new method is proposed that improves on existing methods by automating the identification and classification of lakes, and by introducing a flexible system that can capture the full range of lake forms. Applying our new method we are better able to analyse the evolution of lakes over a number of melt seasons. We find that lakes initiate after approximately 40 positive degree days. Most lakes exist for less than 20 days before draining, or later in the season, and less often, freezing over. Using the automated method developed in this thesis lakes have been mapped in imagery from 2001–2010 at approximately five day intervals. / <p>At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 3: Manuscript. Paper 4: Manuscript. Paper 5: Manuscript.</p>

University of Macau / Faculty of Science and Technology / Department of Mathematics

University of Macau / Faculty of Science and Technology / Department of Mathematics

Active galactic nuclei (AGN) are known to vary in brightness in all regions of the electromagnetic spectrum and over a wide range of timescales. Many methods have been utilized to transform this observed variability into meaningful information about the central engines of AGN. One such technique, adapted from time series analysis of galactic x-ray binary systems, has been used to detect a characteristic break timescale in the power density spectra of x-ray variability in Seyfert galaxies. This timescale, thought to be related to instabilities in the accretion disk, appears to scale with black hole mass over many orders of magnitude. This dissertation performs similar time series analyses with the optical data of eight blazars. The majority of these objects also display a characteristic break timescale. In cases where a black hole estimate is known, the timescales are in good agreement with the relationship observed for galactic x-ray binary systems and Seyfert galaxies. For objects of unknown mass, this relationship can be used to provide a mass estimate of the supermassive black hole. Comparisons are made between the structure function and power density spectrum for each object, and the implications for the connection between the accretion disk and the relativistic jet in AGN are discussed.

This paper explores the relation between the governmental expenditures on the cultural sector and the performance of the economy in the UK. In welfare economies it is the government’s role to shorten the effects of the occurring market failures. It is shown that in the cultural sector, two market failures, the high fix cost and the productivity lag are appearing. In order to ease these effects the government intervening into the market mechanisms by giving grants and subsidies to the stakeholders. In the empirical part a time series analysis is executed between the GDP, the total governmental expenditures and the governmental expenditures on the cultural services on a 60 years interval in the UK. It is shown that the three variables have the same order of integration, they move together over time, furthermore cointegration was detected between them. With Granger causality test it was proven that there is a bidirectional informal connection between the performance of economy and the government’s cultural expenditures.

The number of solar panels deployed worldwide has rapidly increased. Solar panels are often placed in areas not easily accessible. It is also difficult for panel owners to be aware of their operating condition. Many environmental factors have negative effects on the efficiency of solar panels. To reduce the power lost caused by environmental factors, it is necessary to detect and classify the anomalous events occurring on the surface of solar panels. This thesis designs and studies a device to continuously measure the voltage output of solar panels and to transmit the time series data back to a personal computer using wireless communication. A program was developed to store and model this time series data. It also detected the existence of anomalies and classified the anomalies by modeling the data. In total, ten types of anomalies were considered. These anomaly types include temporary shading, permanent shading, fallen leaves, accumulating snow and melting snow among others. Previous time series anomaly detection algorithms do not perform well for reallife situations and are only capable of dealing with at most four different types of anomalies. In this work, a general mathematical model is proposed to give better performance in real-life test cases and to cover more than four types of anomalies. We note that the models can be generalized to detect and to classify anomalies for general time series data which is not necessarily generated from solar panel. We compared several techniques to detect and to classify anomalies including the auto-regressive integrated moving average model (ARIMA), neural networks, support vector machines and k-nearest-neighbors classification. We found that anomaly classification using the k-nearest-neighbors classification was able to accurately detect and classify 97% of the anomalies in our test set. The devices and algorithms have been tested with two small 12-volt solar panels.