Diese Dissertation untersucht einen möglichen Zusammenhang zwischen dem beschriebenen klinischen Verlauf bei Patienten mit konstitutiv aktivierenden TSH-Rezeptormutationen und der gemessenen in vitro Aktivität. Konstitutiv aktivierende Mutationen finden sich als somatische Mutationen in autonomen Adenomen und als Keimbahnmutationen bei Patienten mit sporadischer bzw. familiärer nicht-autoimmuner Hyperthyreose. Die in vitro Aktivität der zu Grunde liegenden TSH-Rezeptormutationen wird mit Hilfe der Linearen Regressions-Analyse bestimmt. Dies ist ein Verfahren, welches die basale Produktion des second messenger cAMP (Cyclo-Adenosinmonophosphat) misst, unter Berücksichtigung der Expression des TSH-Rezeptors. Die Analyse der Krankheitsverläufe der sporadischen nicht-autoimmunen Hyperthyreose zeigt keinen eindeutigen Bezug zur gemessenen in vitro Aktivität. Es besteht jedoch eine höhere in vitro Aktivität bei Mutationen, die sowohl bei der nicht-autoimmunen sporadischen Hyperthyreose und in autonomen Adenomen zu finden sind, im Vergleich zu ausschließlich familiären Mutationen. Dies entspricht auch dem klinischen Eindruck. Für die wenigen bekannten Fälle der sporadischen nicht-autoimmunen Hyperthyreose wurden dramatische Verläufe mit häufigen Rückfällen unter medikamentöser Therapie und zahlreichen Komplikationen (z.B. mentale Retardierung, Kraniosynostose, zerebrale Ventrikulomegalie, beschleunigte Knochenreifung) beschrieben.

In this thesis, we address the identification of biomarkers in high-dimensional omics data. The identification of valid biomarkers is especially relevant for personalized medicine that depends on accurate prediction rules. Moreover, biomarkers elucidate the provenance of disease, or molecular changes related to disease. From a statistical point of view the identification of biomarkers is best cast as variable selection. In particular, we refer to variables as the molecular attributes under investigation, e.g. genes, genetic variation, or metabolites; and we refer to observations as the specific samples whose attributes we investigate, e.g. patients and controls. Variable selection in high-dimensional omics data is a complicated challenge due to the characteristic structure of omics data. For one, omics data is high-dimensional, comprising cellular information in unprecedented details. Moreover, there is an intricate correlation structure among the variables due to e.g internal cellular regulation, or external, latent factors. Variable selection for uncorrelated data is well established. In contrast, there is no consensus on how to approach variable selection under correlation. Here, we introduce a multivariate framework for variable selection that explicitly accounts for the correlation among markers. In particular, we present two novel quantities for variable importance: the correlation-adjusted t (CAT) score for classification, and the correlation-adjusted (marginal) correlation (CAR) score for regression. The CAT score is defined as the Mahalanobis-decorrelated t-score vector, and the CAR score as the Mahalanobis-decorrelated correlation between the predictor variables and the outcome. We derive the CAT and CAR score from a predictive point of view in linear discriminant analysis and regression; both quantities assess the weight of a decorrelated and standardized variable on the prediction rule. Furthermore, we discuss properties of both scores and relations to established quantities. Above all, the CAT score decomposes Hotelling’s T 2 and the CAR score the proportion of variance explained. Notably, the decomposition of total variance into explained and unexplained variance in the linear model can be rewritten in terms of CAR scores. To render our approach applicable on high-dimensional omics data we devise an efficient algorithm for shrinkage estimates of the CAT and CAR score. Subsequently, we conduct extensive simulation studies to investigate the performance of our novel approaches in ranking and prediction under correlation. Here, CAT and CAR scores consistently improve over marginal approaches in terms of more true positives selected and a lower model error. Finally, we illustrate the application of CAT and CAR score on real omics data. In particular, we analyze genomics, transcriptomics, and metabolomics data. We ascertain that CAT and CAR score are competitive or outperform state of the art techniques in terms of true positives detected and prediction error.

In this thesis, we describe two approaches to automatically synthesize the emotional speech of a target speaker based on the hidden Markov model for his/her neutral speech. In the interpolation based method, the basic idea is the model interpolation between the neutral model of the target speaker and an emotional model selected from a candidate pool. Both the interpolation model selection and the interpolation weight computation are determined based on a model-distance measure. We propose a monophone-based Mahalanobis distance (MBMD). In the parallel model combination (PMC) based method, our basic idea is to model the mismatch between neutral model and emotional model. We train linear regression model to describe this mismatch. And then we combine the target speaker neutral model with the linear regression model. We evaluate our approach on the synthesized emotional speech of angriness, happiness, and sadness with several subjective tests. Experimental results show that the implemented system is able to synthesize speech with emotional expressiveness of the target speaker.

In this thesis, a data collection application based on MapReduce programming is described. This application aims to collect tempera- ture data stream continuously from a specied set of sensors. Instead of collecting the temperature information of all the sensors by one machine, the sensors are divided into several subsets each of which is handled as a Map task. In each Map task, the temperature data stream of the assigned sensors is collected continuously and stored in a predened database. All the Map tasks can run simultaneously on several machines. This method can reduce the delay time and improve the eciency of the data collection service, especially in the case of having a huge number of sensors monitored remotely by a data center through Internet. We can use the value of remote sensors to predict the next value of remote sensors by some methods such as linear regres- sion and K-means. And, we can use it to predict the system alarm. Experimental results show that the proposed method is eective in temperature data collection,and eective in carbon reduction.

This dissertation includes two parts. Part 1 develops a geostatistical method to calibrate Texas NexRad rainfall estimates using rain gauge measurements. Part 2 explores the asymptotic joint distribution of sample space-time covariance estimators. The following two paragraphs briefly summarize these two parts, respectively. Rainfall is one of the most important hydrologic model inputs and is considered a random process in time and space. Rain gauges generally provide good quality data; however, they are usually too sparse to capture the spatial variability. Radar estimates provide a better spatial representation of rainfall patterns, but they are subject to substantial biases. Our calibration of radar estimates, using gauge data, takes season, rainfall type and rainfall amount into account, and is accomplished via a combination of threshold estimation, bias reduction, regression techniques and geostatistical procedures. We explore a varying-coefficient model to adapt to the temporal variability of rainfall. The methods are illustrated using Texas rainfall data in 2003, which includes WAR-88D radar-reflectivity data and the corresponding rain gauge measurements. Simulation experiments are carried out to evaluate the accuracy of our methodology. The superiority of the proposed method lies in estimating total rainfall as well as point rainfall amount. We study the asymptotic joint distribution of sample space-time covariance esti-mators of stationary random fields. We do this without any marginal or joint distri-butional assumptions other than mild moment and mixing conditions. We consider several situations depending on whether the observations are regularly or irregularly spaced, and whether one part or the whole domain of interest is fixed or increasing. A simulation experiment illustrates the asymptotic joint normality and the asymp- totic covariance matrix of sample space-time covariance estimators as derived. An extension of this part develops a nonparametric test for full symmetry, separability, Taylor's hypothesis and isotropy of space-time covariances.

ABSTRACT PREDICTION OF LITIGATION PROBABILITY FOR INTERNATIONAL CONSTRUCTION PROJECTS DURING BIDDING STAGE Ayten, ilkay M.S., Department of Civil Engineering Supervisor: Assoc. Prof. Dr. Rifat S&ouml / nmez February 2010, 102 pages Over the years many researchers agreed that between the parties involved in construction projects such as / owner, contractor, engineer and suppliers trying to perform different scopes in different timetables. Therefore, disputes are inevitable due to the complexity of the work. Occurrence of litigation is the most terrifying process to deal with during any construction project for both owner and the contractor because of the time and money consuming nature of the process. Hence, contractors should try to eliminate any potential risk factors that will lead to litigation. The aim of this study is to investigate the factors that influence court action between parties in international construction projects and develop a statistical model that will predict the litigation probability of an international construction project during bidding stage. The final prediction model revealed that contractual awareness and consciousness of risk factors is the key to predict litigation probability. Considering awareness of the factors affecting litigation probability are displayed in this thesis. Companies may have the opportunity to develop risk assessment and management strategies while reconsidering their contingency estimates.

Sound wave has much better transmission in ocean environment than electromagnetic waves, therefore sonar systems are widely applied in underwater investigations. However, not only the target signal is received by the sonar but also the noise from different directions. The noise will affect the performance of the sonar, so the understanding of ocean ambient is an important issue both in academic study and military applications. The ambient noise data of this research was collected by a passive acoustic recording system deployed in the southwest sea of Taiwan, along with the information of wind velocity in the experimented area. The influence on noise level fluctuations by the variation of the wind velocity was first discussed in light of correlation analysis. The fluctuations were expressed in terms of statistic distribution, mean value, standard deviation in different time series. As results, 500 Hz and 1.5k Hz were saturated by high levels signal from unknown sources in spring and summer, so the average sound levels were higher than in fall and winter, about 10 dB and 5 dB higher for 500 Hz and 1.5k Hz respectively. In seasonal analysis, 2.4k and 3.6k Hz have quite stable the mean levels and their standard deviations were around 3 dB. Especially, the noise level of 3.6 Hz has the least fluctuation throughout the year than any other frequencies analyzed. It was also observed that the noise level was decreased with the increase of frequency. Calculated by linear regression, this research worked out the estimation equation for the ambient noise level at high wind speed. However, the estimated values are higher than the measured data, it is due to the distribution of wind velocity. The wind data in this study was skewed towards the lower velocity, consequently the predicted values were overestimated.

Pyrolysis has been a useful procedure to treat waste-tire, which decomposes waste-tire at high temperature in the absence of oxygen. This thermal decomposition process generates pyrolysis oil, combustible gas, and char, which distribute in liquid phase, gas phase, and solid phase, respectively. Pyrolysis oil and combustible gas are fuels, while char is composed of carbon black and ash. Thus, char would be economically worth while to be treated before reuse. In this study, based on the resistivity difference between carbon black and ash, ash can be removed from char in the principle of electrostatic separation and thus increase the value of char. In this study, the objective was to separate ash from char by electrostatic separation process, different char including waste-tire pyrolytic char (raw char), low pressure re-pyrolytic char, ZnO-added char (12% ZnO mixed with 600 oC re-pyrolytic cahr) and man-made char (N600 carbon black mixed with 14.5% metallic oxide) were tested. The Electro-Static Separator (ESS) was designed and constructed with two types of discharge electrodes including a needle-plate electrode (NPE) and a needle-bar electrode (NBE) and two kinds of dust feeders to generate either fine or coarse particles. The results indicated that raw char had higher collection efficiency using the NBE system than the NPE system in the operating voltages of -7 kV to -15 kV because the surface area of the NBE system was less than the NPE system, thus led higher surface charge density for the NBE system than the NPE system, resulting in higher discharge current of the NBE system. In order to lower resistivity and reduce deposited pyrolysis oil on char, low pressure repyrolysis process was used. Because the removal efficiency of pyrolysis oil is proportional to repyrolysis temperature, more pyrolysis oil can be removed from the surface of char, resulting in more carbon blacks exposed on the char surface as conductive material. Thus, the collection efficiency of 600 oC repyrolytic char was less than that of 400 oC repyrolytic char. Furthermore, because particle charging quantity was proportional to particle size, fine char particles had less collection efficiency than coarse char particles. However, both raw char and repyrolytic char, the collection efficiency of carbon and ash had similar trends, suggesting that similar percentage of carbon and ash were collected on the plate and penetrated the ESS system. Therefore, the separation efficiency of carbon and ash were similar, same situation was observed for the ZnO-added char. In order to verify the feasibility of carbon and ash separation by electrostatic separation process, N660 carbon black mixing with 14.5% man-made ash (Al2O3, ZnO and CaO composed) to simulate man-made char, which was further used to proceed the electrostatic separation experiments in this study. The results indicated that the collection efficiency of man-made char increased with operating voltage, and the ash content seems to increase with voltage. Carbon black is a low resistivity material, which causing sparkover during the experiments, thus operating voltage cannot be regulated more than -8.25 kV. In order to verify the feasibility of carbon black and ash separation by the principle of electrostatic separation, this study applied non-linear regression to model the collection efficiency of man-made char, carbon black and ash, and further simulate the collection efficiency at higher electrical field strength. The simulated results indicated that the maximum collection efficiency of carbon and ash was approached around -10 kV/cm of carbon black and ash and their collection efficiencies were similar. The collection efficiency of ash was close to the ash content of man-made char (the collection efficiency of ash equal to the collected ash per mass of injected char), suggesting that most injected ash was collected by the ESS system. In addition, the ash content of penetration char was also simulated, the modeling results showed that the ash content of penetrated char were lower than 2%, while was relatively lower than the raw man-made char, and more than 75% injected char could penetrate the ESS system during the operation procedure. According to the modeling results, solid-solid separation technology could be more efficient if carbon and ash are independently separate particles, and lower resistivity materials would penetrate the ESS system and higher resistivity materials would be collected by the electrostatic separation process.

Fractal was first used in measuring the length of the coastline, with the fractal research and development, not only to break the traditional Archimedean geometry, but also to explain many scientific to ignore the complexity and nature of nonlinear phenomena structure .Fractal has been widely applied to such as physics, astronomy, geography and sociology and other fields, as a wave of interdisciplinary research in recent years. Coastal areas has always been cultural, economic and activities areas since ancient times. Coastal zone was land and sea for the interaction region by a variety of factors (ex: waves, tides, currents and wind, etc.) continue to function, derived from different coastal terrain. Therefore changes in the coast of the deep impact of humanity. Under the principle of the conservation and development, Coastal areas should be use of modern technology to prediction, analysis, assessment, planning, and management, so that a sustainable preservation of coastal resources. In this study, static and dynamic predict and simulation the coast shape base on fractal. The static part is observation of 29 beaches in South China coast. And collect and calculate the parameters and fractal dimensions of the coast. Through the shape of image processing and analysis of information, to find two generators of the coast. Through the data mining technology to identify the criteria for classification, and to simulation the coastline by generate iterations method. The dynamic part is based on hydraulic model¡¦s results, the use of traditional multiple linear regression and neural network to compare the dynamic prediction of the coastline. The results show that the use of neural networks to predict than the use of multiple linear regression, and effect of use difference angle (£c) to predict sub-coastlines than the effect of not use difference angle (£c) to predict, and add fractal dimension can effectively reduce the predict error and increase the degree of interpretation.

During a rod extrusion process, the oxidation layer and the segregation layer at the billet surface are possibly drawn inside the billet and become one part of the product, which portion with surface-permeation has to be cut off and results in a low productivity of the extrusion process. In this paper, the mechanism of permeation of the oxidation and segregation layers at the billet surface is explored using a finite element analysis. The effects of various extrusion conditions, such as extrusion ratio, inclination angle, billet length, the thickness of oxidation layer, etc., on the length of the portion with surface permeation are discussed systematically. Optimal inclination angles for a free surface-permeation product under different extrusion ratios are found out. An empirical equation for the optimal inclination angles is also proposed. Finally, experiments of extrusion of aluminum rods are conducted to validate the analytical model proposed.