The mobile traffic in cellular networks is increasing in a steady rate as we go intoa future where we are connected to the internet practically all the time in one wayor another. To map the mobile traffic and the volume pressure on the base stationduring different time periods, it is useful to have the ability to predict the trafficvolumes within cellular networks. The data in this work consists of 4G cellular trafficdata spanning over a 7 day coherent period, collected from cells in a moderately largecity. The proposed method in this work is ARIMA modeling, in both original formand with an extension where the coefficients of the ARIMA model are re-esimated byintroducing some user characteristic variables. The re-estimated coefficients produceslightly lower forecast errors in general than a isolated ARIMA model where thevolume forecasts only depends on time. This implies that the forecasts can besomewhat improved when we allow the influence of these variables to be a part ofthe model, and not only the time series itself.

Avian influenza H5N1 is one kind of important bird flu. Unfortunately, this virus has swiftly evolved and become highly pathogenic to humans and poultry, resulting in 100% of death in infected poultry and over 60% of mortality among infected human population. Moreover, the virus tends to reassort with other influenza viruses, such as the current swine flu H1N1, to establish themselves in environments and further this epidemic all over the world. The World Health Organization (WHO) has in fact warned that highly pathogenic avian influenza H5N1 poses a graver risk of a global human pandemic than at any time since the Hong Kong outbreak (H3N2) in the 1960s. / Finally, avian influenza is an inter-disciplinary issue across virology, medical geography, and spatial epidemiology. How to quantify and integrate knowledge from different disciplines remains a challenge in fully understanding the disease. We propose a method to formally integrate genetic analysis that identifies the evolution of the H5N1 virus in space and time, epidemiological analysis that determines socio-environmental factors associated with H5N1 occurrence and statistical analysis that identifies outbreak dusters. Our integrated results show a significant advance in findings over reports in, for instance, Gilbert et al. (2008) and we believe our findings are more precise and informative in representing the occurrence and the space-time dynamics of H5N1 spread. Overall, unlike traditional influenza studies, our work sets up a solid foundation for the inter-disciplinary study of this and other spatial infectious diseases. / First, we apply multifractal detrended fluctuation analysis to determine the temporal scaling behavior of outbreaks in Asia, Europe, Africa, and the whole of the world between December 2003 to March 2009. Long-range correlation and multifractality, two important properties characterizing the scaling behavior of complex dynamics, are first detected in the outbreak time series. In addition, this study identifies different temporal scaling behaviors of outbreaks of these continents 8,nd specific seasonal patterns in Asia. These findings confirm our perspective that avian-influenza outbreak behaviors are self-similar over time and are spatially heterogeneous. / One key to preventing such a calamity is to obtain a thorough understanding of the mechanisms of avian influenza transmission and its spatio-temporal patterns of dispersal. The issues at stake are outbreaks' spatial and temporal patterns, the interrelationship of these with the evolution of influenza viruses in such a way that geography is understood as a dimension of the disease's virology, and the human and avian behaviors and socio-ecological environments associated with H5Nl spread. This thesis sets out to study these problems in detail and propose solutions. / Second, we conduct a spatial analysis for global trends and local clusters of H5N1 outbreaks at multiple geographical scales. Currently, the local K function used in a point pattern analysis searches outbreak clusters, assuming the disease is spatially homogeneous. The thesis proposes a much more efficient method to measure the degree of clusters accurately. The modified function works by weighting outbreaks through distances, counting the number of the weighted outbreaks for each lattice point no matter whether the disease emerges in a grid. This weighted local K function extends cluster analysis from a point pattern to lattice data. Spatial representation in these terms then seeks to explore local patterns of H5N1 over a continuous space. / Third, we study a set of socio-environmental factors, which are plausibly associated with the occurrence of H5N1. Spatial epidemiological models are built for predicting the disease at both continental and national levels, covering Indonesia, China, and the whole of East-Southeast Asia. We evaluate the statistical models using 1,000 bootstrap replicates, showing a consistently high rate of prediction, assessed by statistics: AUC, Kappa Index, and pseudo R square. / Ge, Erjia. / Advisers: Yee Leung; Tung Fung. / Source: Dissertation Abstracts International, Volume: 73-06, Section: A, page: . / Thesis (Ph.D.)--Chinese University of Hong Kong, 2011. / Includes bibliographical references (leaves 169-197). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [201-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese.

Social media has created new protest spaces and has enabled people to do things differently. The focus of the research is on campaign groups, created before social media was used as a tool for protest. It has been undertaken to achieve the aim of the challenges and opportunities of incorporating new forms of social media into existing protest campaigns through a case study of anti-airport expansion groups in the UK. Social media data was obtained from three anti-airport expansion groups which included the extraction of approximately 9,000 tweets and 8,000 Facebook posts. The data were then analysed using social network analysis, time series analysis and semi-structured interviews. The results of social network analysis and time series analysis informed the development of the questions directed at the social media coordinators of each group. The main findings are that Airport Watch and HACAN Clearskies exhibit very similar Twitter networks and favour interaction with the media, similar anti-airport expansion groups and also pro-airport expansion groups. Transition Heathrow demonstrates more varied interaction patterns. All groups dominate their respective Facebook page and group networks apart from HACAN Clearskies which has non-assigned leaders controlling information dissemination in the group. Time series analysis uncovered variations in social media usage; overall for all three campaign groups Twitter was utilised more than Facebook.

Objective: One of the main theoretical proposals of Intensive Short-term Dynamic Psychotherapy (ISTDP; Davanloo, 1990) is that experiencing of previously unconscious guilt over aggressive impulses associated with attachment trauma leads to increase in self-compassion. The present study aimed to test this assumption. Method: Videotaped sessions from five therapies from a randomized controlled trial of 20-sessions of time-limited ISTDP for treatment-refractory depression were rated with the Achievement of Therapeutic Objectives Scale (ATOS; McCullough, Larsen, Schanche, Andrews& Kuhn, 2003b). Degree of patient guilt arousal and self-compassion were rated on all available sessions. Data were analyzed using a replicated single-subject time-series approach. Results: Guilt arousal was not shown to positively predict self-compassion for any of the five patients. For one patient guilt arousal negatively predicted self-compassion two sessions ahead in time. Conclusion: The current study yields no support that the experience of guilt over aggressive feelings and impulses leads to increases in self-compassion. On the contrary, the finding that guilt negatively predicted self-compassion for one patient must be considered as an indication that this treatment process might negatively impact self-compassion for some patients in some contexts. However, there are several methodological limitations to the current study in the light of which the results should be regarded as tentative.

Due to its generality and flexibility, the state-space model has become one of the most popular models in modern time domain analysis for the description and prediction of time series data. The model is often used to characterize processes that can be conceptualized as "signal plus noise," where the realized series is viewed as the manifestation of a latent signal that has been corrupted by observation noise. In the state-space framework, parameter estimation is generally accomplished by maximizing the innovations Gaussian log-likelihood. The maximum likelihood estimator (MLE) is efficient when the normality assumption is satisfied. However, in the presence of contamination, the MLE suffers from a lack of robustness. Basu, Harris, Hjort, and Jones (1998) introduced a discrepancy measure (BHHJ) with a non-negative tuning parameter that regulates the trade-off between robustness and efficiency. In this manuscript, we propose a new parameter estimation procedure based on the BHHJ discrepancy for fitting state-space models. As the tuning parameter is increased, the estimation procedure becomes more robust but less efficient. We investigate the performance of the procedure in an illustrative simulation study. In addition, we propose a numerical method to approximate the asymptotic variance of the estimator, and we provide an approach for choosing an appropriate tuning parameter in practice. We justify these procedures theoretically and investigate their efficacy in simulation studies. Based on the proposed parameter estimation procedure, we then develop a new model selection criterion in the state-space framework. The traditional Akaike information criterion (AIC), where the goodness-of-fit is assessed by the empirical log-likelihood, is not robust to outliers. Our new criterion is comprised of a goodness-of-fit term based on the empirical BHHJ discrepancy, and a penalty term based on both the tuning parameter and the dimension of the candidate model. We present a comprehensive simulation study to investigate the performance of the new criterion. In instances where the time series data is contaminated, our proposed model selection criterion is shown to perform favorably relative to AIC. Lastly, using the BHHJ discrepancy based on the chosen tuning parameter, we propose two versions of an influence diagnostic in the state-space framework. Specifically, our diagnostics help to identify cases that influence the recovery of the latent signal, thereby providing initial guidance and insight for further exploration. We illustrate the behavior of these measures in a simulation study.

Abstract not available

This thesis describes a new approach to achieving high quality distributed garbage collection using the Train Algorithm. This algorithm has been investigated for its ability to provide high quality collection in a variety of contexts, including persistent object systems and distributed object systems. Prior literature on the distributed Train Algorithm suggests that safe, complete, asynchronous, and scalable collection can be attained, however an approach that achieves this combination of behaviour has yet to emerge. The mechanisms and policies described in this thesis are unique in their ability to exploit the distributed Train Algorithm in a manner that displays all four desirable qualities. Further the mechanisms allow any number of mutator and collector threads to operate concurrently within a site; this is also a unique property amongst train-based mechanisms (distributed or otherwise). Confidence in the quality of the approach promoted in this thesis is obtained via a top-down approach. Firstly a concise behavioural model is introduced to capture fundamental requirements for safe and complete behaviour from train-based collection mechanisms. The model abstracts over the techniques previously introduced under the banner of the Train Algorithm. It serves as a self- contained template for correct train-based collection that is independent of a target object system for deployment of the algorithm. Secondly a means to instantiate the model in a distributed object system is described. The instantiation includes well-established techniques from prior literature, and via the model these are correctly refined and reorganised with new techniques to achieve asynchrony, scalability, and support for concurrency. The result is a flexible approach that allows a distributed system to exhibit a variety of local collection mechanisms and policies, while ensuring their interaction is safe, complete, asynchronous, and scalable regardless of the local choices made by each site. Additional confidence in the properties of the new approach is obtained from implementation within a distributed object system simulation. The implementation provides some insight into the practical issues that arise through the combination of distribution, concurrent execution within sites, and train-based collection. Executions of the simulation system are used to verify that safe collection is observed at all times, and obtain evidence that asynchrony, scalability, and concurrency can be observed in practice. / Thesis (Ph.D.)--School of Computer Science, 2004.

Signal detection from Adverse Event Reports (AERs) is important for identifying and analysing drug safety concern after a drug has been released into the market. A safety signal is defined as a possible causal relation between an adverse event and a drug. There are a number of safety signal detection algorithms available for detecting drug safety concern. They compare the ratio of observed count to expected count to find instances of disproportionate reportings of an event for a drug or combination of events for a drug. In this thesis, we present an algorithm to mine the AERs to identify drugs which show sudden and large changes in patterns of reporting of adverse events. Unlike other algorithms, the proposed algorithm creates time series for each drug and use it to identify start of a potential safety problem. A novel vectorized timeseries utilizing multiple attributes has been proposed here. First a time series with a small time period was created; then to remove local variations of the number of reports in a time period, a time-window based averaging was done. This method helped to keep a relatively long time-series, but eliminated local variations. The steps in the algorithm include partitioning the counts on attribute values, creating a vector out of the partitioned counts for each time period, use of a sliding time window, normalizing the vectors and computing vector differences to find the changes in reporting over time. Weights have been assigned to attributes to highlight changes in the more significant attributes. The algorithm was tested with Adverse Event Reporting System (AERS) datasets from Food and Drug Administation (FDA). From AERS datasets the proposed algorithm identified five drugs that may have safety concern. After searching literature and the Internet it was found that the five drugs the algorithm identified, two were recalled, one was suspended, one had to undergo label change and the other one has a lawsuit pending against it.

In der vorliegenden Dissertation wird eine Beschreibung der Phasendynamik irregulärer Oszillationen und deren Wechselwirkungen vorgestellt. Hierbei werden chaotische und stochastische Oszillationen autonomer dissipativer Systeme betrachtet. Für eine Phasenbeschreibung stochastischer Oszillationen müssen zum einen unterschiedliche Werte der Phase zueinander in Beziehung gesetzt werden, um ihre Dynamik unabhängig von der gewählten Parametrisierung der Oszillation beschreiben zu können. Zum anderen müssen für stochastische und chaotische Oszillationen diejenigen Systemzustände identifiziert werden, die sich in der gleichen Phase befinden. Im Rahmen dieser Dissertation werden die Werte der Phase über eine gemittelte Phasengeschwindigkeitsfunktion miteinander in Beziehung gesetzt. Für stochastische Oszillationen sind jedoch verschiedene Definitionen der mittleren Geschwindigkeit möglich. Um die Unterschiede der Geschwindigkeitsdefinitionen besser zu verstehen, werden auf ihrer Basis effektive deterministische Modelle der Oszillationen konstruiert. Hierbei zeigt sich, dass die Modelle unterschiedliche Oszillationseigenschaften, wie z. B. die mittlere Frequenz oder die invariante Wahrscheinlichkeitsverteilung, nachahmen. Je nach Anwendung stellt die effektive Phasengeschwindigkeitsfunktion eines speziellen Modells eine zweckmäßige Phasenbeziehung her. Wie anhand einfacher Beispiele erklärt wird, kann so die Theorie der effektiven Phasendynamik auch kontinuierlich und pulsartig wechselwirkende stochastische Oszillationen beschreiben. Weiterhin wird ein Kriterium für die invariante Identifikation von Zuständen gleicher Phase irregulärer Oszillationen zu sogenannten generalisierten Isophasen beschrieben: Die Zustände einer solchen Isophase sollen in ihrer dynamischen Entwicklung ununterscheidbar werden. Für stochastische Oszillationen wird dieses Kriterium in einem mittleren Sinne interpretiert. Wie anhand von Beispielen demonstriert wird, lassen sich so verschiedene Typen stochastischer Oszillationen in einheitlicher Weise auf eine stochastische Phasendynamik reduzieren. Mit Hilfe eines numerischen Algorithmus zur Schätzung der Isophasen aus Daten wird die Anwendbarkeit der Theorie anhand eines Signals regelmäßiger Atmung gezeigt. Weiterhin zeigt sich, dass das Kriterium der Phasenidentifikation für chaotische Oszillationen nur approximativ erfüllt werden kann. Anhand des Rössleroszillators wird der tiefgreifende Zusammenhang zwischen approximativen Isophasen, chaotischer Phasendiffusion und instabilen periodischen Orbits dargelegt. Gemeinsam ermöglichen die Theorien der effektiven Phasendynamik und der generalisierten Isophasen eine umfassende und einheitliche Phasenbeschreibung irregulärer Oszillationen. / Many natural systems embedded in a complex surrounding show irregular oscillatory dynamics. The oscillations can be parameterized by a phase variable in order to obtain a simplified theoretical description of the dynamics. Importantly, a phase description can be easily extended to describe the interactions of the system with its surrounding. It is desirable to define an invariant phase that is independent of the observable or the arbitrary parameterization, in order to make, for example, the phase characteristics obtained from different experiments comparable. In this thesis, we present an invariant phase description of irregular oscillations and their interactions with the surrounding. The description is applicable to stochastic and chaotic irregular oscillations of autonomous dissipative systems. For this it is necessary to interrelate different phase values in order to allow for a parameterization-independent phase definition. On the other hand, a criterion is needed, that invariantly identifies the system states that are in the same phase. To allow for a parameterization-independent definition of phase, we interrelate different phase values by the phase velocity. However, the treatment of stochastic oscillations is complicated by the fact that different definitions of average velocity are possible. For a better understanding of their differences, we analyse effective deterministic phase models of the oscillations based upon the different velocity definitions. Dependent on the application, a certain effective velocity is suitable for a parameterization-independent phase description. In this way, continuous as well pulse-like interactions of stochastic oscillations can be described, as it is demonstrated with simple examples. On the other hand, an invariant criterion of identification is proposed that generalizes the concept of standard (Winfree) isophases. System states of the same phase are identified to belong to the same generalized isophase using the following invariant criterion: All states of an isophase shall become indistinguishable in the course of time. The criterion is interpreted in an average sense for stochastic oscillations. It allows for a unified treatment of different types of stochastic oscillations. Using a numerical estimation algorithm of isophases, the applicability of the theory is demonstrated by a signal of regular human respiration. For chaotic oscillations, generalized isophases can only be obtained up to a certain approximation. The intimate relationship between these approximate isophase, chaotic phase diffusion, and unstable periodic orbits is explained with the example of the chaotic roes oscillator. Together, the concept of generalized isophases and the effective phase theory allow for a unified, and invariant phase description of stochastic and chaotic irregular oscillations.

The increasing amount of information makes sequential data mining an important domain of research. A vast number of data mining models and approaches have been developed in order to extract interesting and useful patterns of data. Most models are used for strategic purposes resulting in using of the time parameter. However, the extensive field of data mining applications requires new models to be introduced. The current thesis proposed models for temporal sequential data mining having as a goal the forecasting process. We focus our study on sequential temporal database analysis and on time-series data. In sequential database analysis we propose several interestingness measures for rules selection and patterns extraction. Their goal is to advantage those rules/patterns whose time-distance between the itemsets is small. The extracted information is used to predict user¿s future requests in a web log database, obtaining a higher performance in comparison to other compared models. In time-series analysis we propose a forecasting model based on Neural Networks, Genetic Algorithms, and Wavelet Transform. We apply it on a WiMAX network traffic and EUR/USD currency exchange data in order to compare its prediction performance with those obtained using other existing models. Different ways of changing parameters adapted to a given situation and the corresponding simulations are presented. It was shown that the proposed model outperforms the existing ones from the prediction point of view on the used time-series. As a whole, this thesis proposes forecasting models for different types of temporal sequential data with different characteristics and behaviour.