No description available.

No description available.

Parameter estimation is a long-lasting topic in queueing systems and has attracted considerable attention from both academia and industry. In this thesis, we design a parameter estimation framework for a tandem queueing system that collects end-to-end measurement data and utilizes the finite mixture model for the maximum likelihood (ML) estimation. The likelihood equations produced by ML are then solved by the iterative expectation-maximization (EM) algorithm, a powerful algorithm for parameter estimation in scenarios involving complicated distributions. We carry out a set of experiments with different parameter settings to test the performance of the proposed framework. Experimental results show that our method performs well for tandem queueing systems, in which the constituent nodes' service time follow distributions governed by exponential family. Under this framework, both the Newton-Raphson (NR) algorithm and the EM algorithm could be applied. The EM algorithm, however, is recommended due to its ease of implementation and lower computational overhead. / Graduate / hangli@uvic.ca

We evaluate the security of human voice password databases from an information theoretical point of view. More specifically, we provide a theoretical estimation on the amount of entropy in human voice when processed using the conventional GMM-UBM technologies and the MFCCs as the acoustic features. The theoretical estimation gives rise to a methodology for analyzing the security level in a corpus of human voice. That is, given a database containing speech signals, we provide a method for estimating the relative entropy (Kullback-Leibler divergence) of the database thereby establishing the security level of the speaker verification system. To demonstrate this, we analyze the YOHO database, a corpus of voice samples collected from 138 speakers and show that the amount of entropy extracted is less than 14-bits. We also present a practical attack that succeeds in impersonating the voice of any speaker within the corpus with a 98% success probability with as little as 9 trials. The attack will still succeed with a rate of 62.50% if 4 attempts are permitted. Further, based on the same attack rationale, we mount an attack on the ALIZE speaker verification system. We show through experimentation that the attacker can impersonate any user in the database of 69 people with about 25% success rate with only 5 trials. The success rate can achieve more than 50% by increasing the allowed authentication attempts to 20. Finally, when the practical attack is cast in terms of an entropy metric, we find that the theoretical entropy estimate almost perfectly predicts the success rate of the practical attack, giving further credence to the theoretical model and the associated entropy estimation technique.

Analysis of patent documents is one way to learn about trends in the evolutionof technologies. In this thesis, we propose a mixture of life cycle Poisson modelfor predicting the life cycle of technologies from patent count data. The aim is topredict the life cycle of technologies and determine the stage of the technology inthe development S-curve. The model is constructed from historical data on patentpublications of technologies and also from experts’ belief of life cycle of technologies. The methods used to estimate the model are based on Bayesian methods, inparticular we use a combination of Gibbs sampling and slice sampling to simulatefrom the posterior distribution of the model parameters. We apply the model on adataset of 123 technologies from the electricity sector. As a preliminary exploratorystep clustering analysis is also applied on the dataset. Finally we evaluate the modelhow it performs to predict the trend of life cycle of technologies based on differentbase years. Results reveal that the model is capable of predicting the life cycleof technologies based on its different stages. However, the predictions of expectedbehavior become more accurate when more data is used to construct the prediction.

<p>Communication over today’s IP-based networks are to some extent subject to packet loss. Most real-time applications, such as video streaming, need methods to hide this effect, since resending lost packets may introduce unacceptable delays. For IP-based video streaming applications such a method is referred to as a <em>packet loss concealment </em>scheme.</p><p>In this thesis a recently proposed mixture model and least squares-based packet loss concealment scheme is implemented and evaluated together with three more well known concealment methods. The JM reference software is used as basis for the implementation, which is a public available software codec for the H.264 video coding standard. The evaluation is carried out by comparing the schemes in terms of objective measurements, subjective observations and a study with human observers.</p><p>The recently proposed packet loss concealment scheme shows good performance with respect to the objective measures, and careful observations indicate better concealment of scenes with fast motion and rapidly changing video content. The study with human observers verifies the results for the case when a more sophisticated packetization technique is used.</p><p>A new packet loss concealment scheme, based on joint modeling of motion vectors and pixels, is also investigated in the last chapter as an additional contribution of the thesis.</p>

No description available.

In the past decade, the availability of digitally encoded, downloadable music has increased dramatically, pushed mainly by the release of the now famous MP3 compression format (Fraunhofer-Gesellschaft, 1994). Online sales of music in the US doubled in 2005, according to a recent news article (*), while the number of files exchanged on P2P platforms is much higher, but hard to estimate. The existing and coming informational flood in digital music prompts the need for sophisticated content-based information retrieval. Query-by-Humming is a prototypical technique aimed at locating pieces of music by melody; automatic annotation algorithms seek to enable finer search criteria, such as instruments, genre, or meter. Score transcription systems strive for an abstract, compressed form of a piece of music understandable by composers and musicians. Much research still has to be performed to achieve these goals. This thesis connects essential knowledge about music and human auditory perception with signal processing algorithms to solve the specific problem of pitch estimation. The designed algorithm obtains an estimate of the magnitude spectrum via STFT and models the harmonic structure of each pitch contained in the magnitude spectrum with Gaussian density mixtures, whose parameters are subsequently estimated via an Expectation-Maximization (EM) algorithm. Heuristics for EM initialization are formulated mathematically. The system is implemented in MATLAB, featuring a GUI that provides for visual (spectrogram) and numerical (console) verification of results. The algorithm is tested using an array of data ranging from single to triple superposed instrument recordings. Its advantages and limitations are discussed, and a brief outlook over potential future research is given. (*) "Online and Wireless Music Sales Tripled in 2005"; Associated Press; January 19, 2006

The number of longitudinal studies has increased steadily in various social science disciplines over the last decade. Growth Mixture Modeling (GMM) has emerged among the new approaches for analyzing longitudinal data. It can be viewed as a combination of Hierarchical Linear Modeling, Latent Growth Curve Modeling and Finite Mixture Modeling. The combination of both continuous and categorical latent variables makes GMM a flexible analysis procedure. However, when researchers analyze their data using GMM, some may assume that the units are independent of each other even though it may not always be the case. The purpose of this dissertation was to examine the impact of ignoring a higher nesting structure in Multilevel Growth Mixture Modeling on the accuracy of classification of individuals and the accuracy on tests of significance (i.e., Type I error rate and statistical power) of the parameter estimates for the model in each subpopulation. Two simulation studies were conducted. In the first study, the impact of misspecifying the multilevel mixture model is investigated by ignoring a level of nesting structure in cross-sectional data. In the second study, longitudinal clustered data (e.g., repeated measures nested within units and units nested within clusters) are analyzed correctly and with a misspecification ignoring the highest level of the nesting structure. Results indicate that ignoring a higher level nesting structure results in lower classification accuracy, less accurate fixed effect estimates, inflation of lower-level variance estimates, and less accurate standard error estimates, the latter result which in turn affects the accuracy of tests of significance for the fixed effects. The magnitude of the intra-class correlation (ICC) coefficient has a substantial impact when a higher level nesting structure is ignored; the higher the ICC, the more variance at the highest level is ignored, and the worse the performance of the model. The implication for applied researchers is that it is important to model the multilevel data structure in (growth) mixture modeling. In addition, researchers should be cautious in interpreting their results if ignoring a higher level nesting structure is inevitable. Limitations concerning appropriate use of latent class analysis in growth modeling include unknown effects of incorrect estimation of the number of latent classes, non-normal distribution effects, and different growth patterns within-group and between-group.

In this thesis, the speaker and emotion recognition system is established by PC and digit signal processor (DSP). Most speaker and emotion recognition systems are separately accomplished, but not combined together in the same system. In this thesis, it will show how speaker and emotion recognition systems are combined in the same system. In this system, the voice is picked up by a mike and through DSP to extract the characteristics. Then it passes the sample correctly, it can draw the result of distinguishing. The recognition system is divided into four sub-systems: the pronunciation pre-process, the speaker training model, the speaker and emotion recognition, and the speaker confirmation. The pronunciation pre-process uses the mike to capture the voice, and through the DSP board to convey the voice to the SRAM, then movements dealt with pre-process. The speaker trained model uses the Gaussian mixture model to establish the average, coefficient of variation and weight value of the person who sets up speaker specifically. And we¡¦ll take this information to be the datum of the whole recognition system. The speaker recognition mainly uses the density of probability to recognition the speaker¡¦s identity. The emotion recognition takes advantage of the coefficient of variation to recognize the emotion. The speaker confirms is set up to sure whether the user is the same speaker who hits for the systematic database. The recognition system based on DSP includes two parts¡GHardware setting and implementation of speaker algorithm. We use the fixed-arithmetician DSP chip (chipboard) in the DSP, the algorithm of recognition is Gaussian mixture model. In addition, compared with floating point, the fixed point DSP cost much less; it makes the system nearer to users.