Vid analys av fluider med akustisk spektroskopi finns ett behov av att finna multivariata metoder för att utifrån akustiska spektra prediktera storheter såsom viskositet och densitet. Användning av artificiella neurala nätverk och bayesiska nätverk för detta syfte utreds genom teoretiska och praktiska undersökningar. Förbehandling och uppdelning av data samt en handfull linjära och olinjära multivariata analysmetoder beskrivs och implementeras. Prediktionsfelen för de olika metoderna jämförs och PLS (Partial Least Squares) framstår som den starkaste kandidaten för att prediktera de sökta storheterna. / When analyzing fluids using acoustic spectrometry there is a need of finding multivariate methods for predicting properties such as viscosity and density from acoustic spectra. The utilization of artificial neural networks and Bayesian networks for this purpose is analyzed through theoretical and practical investigations. Preprocessing and division of data along with a handful of linear and non-linear multivariate methods of analysis are described and implemented. The errors of prediction for the different methods are compared and PLS (Partial Least Squares) appear to be the strongest candidate for predicting the sought-after properties.

The emergence of multi-band communications standards, and the fast pace of the consumer electronics markets for wireless/cellular applications emphasize the need for fast design closure. In addition, there is a need for electronic product designers to collaborate with manufacturers, gain essential knowledge regarding the manufacturing facilities and the processes, and apply this knowledge during the design process. In this dissertation, efficient layout-level circuit sizing techniques, and methodologies for design-for-manufacturability have been investigated. For cost-effective fabrication of RF modules on emerging technologies, there is a clear need for design cycle time reduction of passive and active RF modules. This is important since new technologies lack extensive design libraries and layout-level electromagnetic (EM) optimization of RF circuits become the major bottleneck for reduced design time. In addition, the design of multi-band RF circuits requires precise control of design specifications that are partially satisfied due to manufacturing variations, resulting in yield loss. In this work, a broadband modeling and a layout-level sizing technique for embedded inductors/capacitors in multilayer substrate has been presented. The methodology employs artificial neural networks to develop a neuro-model for the embedded passives. Secondly, a layout-level sizing technique for RF passive circuits with quasi-lumped embedded inductors and capacitors has been demonstrated. The sizing technique is based on the circuit augmentation technique and a linear optimization framework. In addition, this dissertation presents a layout-level, multi-domain DFM methodology and yield optimization technique for RF circuits for SOP-based wireless applications. The proposed statistical analysis framework is based on layout segmentation, lumped element modeling, sensitivity analysis, and extraction of probability density functions using convolution methods. The statistical analysis takes into account the effect of thermo-mechanical stress and process variations that are incurred in batch fabrication. Yield enhancement and optimization methods based on joint probability functions and constraint-based convex programming has also been presented. The results in this work have been demonstrated to show good correlation with measurement data.

The performance of radar detection and imaging systems strongly depends on the characteristics of radar clutter. In order to improve the radar signal processing algorithms, successful analysis and modeling of radar clutter are required. For a successful model of radar clutter, both the spectral and statistical characteristics of the clutter should be revealed. Within the scope of this study, an experimental radar data acquisition system is established to analyze radar clutter. The hardware and the data processing system are first verified using generic signals and then a set of measurements is taken in the open terrain. In this thesis, the limitations and problems encountered during the establishment of the system are explained in detail. The spectral and statistical analyses performed on the recorded data are examined. The temporal and spatial behavior of the measured clutter data are explored. The hypothetical models proposed so far in the literature are tested on the experimental data and the fitting of models to the experimental data is confirmed using various goodness-of-fit tests. Finally, the results of the analyses are interpreted in the light of the radar system parameters and the characteristics of the illuminated terrain.

Οι αποθέσεις φλύσχη που εμφανίζονται στο νησί της Καρπάθου μελετήθηκαν από διάφορα σημεία προσέγγισης με σκοπό να εξαχθούν συμπεράσματα σχετικά με χαρακτηριστικά των συγκεκριμένων ιζημάτων τα οποία δεν ήταν γνωστά με λεπτομέρεια μέχρι σήμερα. Το συνολικό πάχος των αποθέσεων δεν ξεπερνά τα 1000 μέτρα και κυμαίνεται σε περίπου 900 για τη βόρεια και περίπου 800 για τη νότια περιοχή εμφάνισης. Όσον αφορά την ιζηματολογία, στρωματογραφία και τα περιβάλλοντα των αποθέσεων του φλύσχη μπορούν να ειπωθούν τα εξής: Τα ιζήματα της βόρειας περιοχής χωρίζονται σε 5 ενότητες, ενώ τα ιζήματα της νότιας περιοχής χωρίζονται σε 4 ενότητες από τη βάση προς την οροφή της ακολουθίας. Η χρονολόγηση των αποθέσεων του φλύσχη με βάση ασβεστιτικά νανοαπολιθώματα υπέδειξε ότι η ιζηματογένεση του φλύσχη στη περιοχή της Καρπάθου φαίνεται να ξεκινάει στο Κατώτερο Ηώκαινο και να εξελίσσεται έως τα όρια Ηωκαίνου-Ολιγοκαίνου. Η στατιστική ανάλυση των παχών στρωμάτων που εφαρμόστηκε έδειξε ότι πρεπει να υπάρχει προσοχή στην ερμηνεία περιβαλλόντων ιζηματογένεσης σε υποθαλάσσια ριπίδια που δίνεται με βάση περιγραφικές τεχνικές. Γεωχημική ανάλυση των φλυσχικών αποθέσεων για την περιεκτικότητα τους σε κύρια στοιχεία, ιχνοστοιχεία και σπάνιες γαίες σε συνδυασμό με πετρογραφική έρευνα έδωσε πολύτιμα στοιχεία για την προέλευση και το γεωτεκτονικό περιβάλλον της λεκάνης ιζηματογένεσης των φλυσχικών ιζημάτων. Η έρευνα για την πιθανή ύπαρξη πεδίου υδρογονανθράκων στις φλυσχικές αποθέσεις της Καρπάθου δεν έδωσε ενθαρρυντικά αποτελέσματα. Με βάση τις παραπάνω παρατηρήσεις και τη γεωλογία της ευρύτερης περιοχής του ΝΑ Αιγαίου και των Δωδεκανήσων, η απόθεση των ιζημάτων του φλύσχη στη Κάρπαθο φαίνεται να έχει άμεση σχέση με τις ορογενετικές κινήσεις των Ταυρίδων οροσειρών κατά το Ηώκαινο-Κατώτερο Ολιγόκαινο. / The flysch deposits that outcrop on the island of Karpathos (SE Greece) were studied, using several approaches, in order to investigate detailed sedimentological characteristics of these deposits which were unknown until now. The flysch deposits outcrop on 2 areas of the island: a northern area (Spoa-Olympos-Diafani) and a southern area (Pigadia-Aperi). The total stratigraphic thickness of the deposits is less than 1000 meters and ranges from about 900 meters for the northern area, to about 800 meters for the southern area. Northern area deposits can be subdivided into 5 sedimentary units and southern area into a sedimentary units (from the bottom to the top of the succession) based on their sedimentological characteristics. Age determination of the flysch based on calcareous nanofossils reveals that flysch sedimentation probably started in Lower Eocene and ended in Upper Eocene-Lower Oligocene. Application of bed thickness statistical analysis revealed that descriptive recognition of sedimentary environments in turbidite deposits is not always reliable. Petrographic and geochemical analysis of the flysch deposits, gave valuable clues regarding the sedimentary source and the tectonic arrangement of the sedimentary basin. Hydrocarbon exploration shows negative results regarding the possible existence of a hydrocarbon field in Karpathos flysch deposits. Taking into account all the above observations and the geological arrangement of the SE Aegean and the Dodecanese Islands, it seems that flysch sedimentation in Karpathos Island has a direct relation with the Taurides orogeny in Eocene-Lower Oligocene times.

Στην εργασία χρησιμοποιούνται σήματα καρδιακού ρυθμού, τα οποία αντιστοιχούν σε φυσιολογικές και οξαιμικές περιπτώσεις.Στην συνέχεια αυτά τα σήματα τα επεξεργαζόμαστε με διάφορες τεχνικές. Στόχος της εργασίας αυτής είναι ο διαχωρισμός των δυο αυτών ομάδων. / In this thesis are used signals of cardiac rythm, that correspond in physiologic and oxidemic cases.Then we processed these signals with various techniques.Target of this thesis is the segregation of this two teams.

No description available.

Surface topography is, generally, composed of many length scales starting from its physical geometry, to its microscopic or atomic scales known by roughness. The spatial and geometrical evolution of the roughness topography of engineering surfaces avail comprehensive understanding, and interpretation of many physical and engineering problems such as friction, and wear mechanisms during the mechanical contact between adjoined surfaces. Obviously, the topography of rough surfaces is of random nature. It is composed of irregular hills/valleys being spatially correlated. The relation between their densities and their geometric properties are the fundamental topics that have been developed, in this research study, using the theory of random fields and the integral geometry.An appropriate random field model of a rough surface has been defined by the most significant parameters, whose changes influence the geometry of its excursion. The excursion sets were quantified by functions known as intrinsic volumes. These functions have many physical interpretations, in practice. It is possible by deriving their analytical formula to estimate the parameters of the random field model being applied on the surface, and for statistical analysis investigation of its excursion sets. These subjects have been essentially considered in this thesis. Firstly, the intrinsic volumes of the excursion sets of a class of mixture models defined by the linear combination of Gaussian and t random fields, then for the skew-t random fields are derived analytically. They have been compared and tested on surfaces generated by simulations. In the second stage, these random fields have been applied to real surfaces measured from the UHMWPE component, involved in application of total hip implant, before and after wear simulation process. The primary results showed that the skew-t random field is more adequate, and flexible for modelling the topographic roughness. Following these arguments, a statistical analysis approach, based on the skew-t random field, is then proposed. It aims at estimating, hierarchically, the significant levels including the real hills/valleys among the uncertain measurements. The evolution of the mean area of the hills/valleys and their levels enabled describing the functional behaviour of the UHMWPE surface over wear time, and indicating the predominant wear mechanisms.

In both developed and emerging-economies, major cities continue to experience increasing traffic congestion. To address this issue, complex Traffic Management Systems (TMS) are employed in recent years to help manage traffic. These systems fuse traffic-surveillance-related information from a variety of sensors deployed across traffic networks. A TMS requires real-time information to make effective control decisions and to deliver trustworthy information to users, such as travel time, congestion level, etc. There are three fundamental inputs required by TMS, namely, traffic volume, vehicular speed, and traffic density. Using conventional traffic loop detectors one can directly measure flow and velocity. However, traffic density is more difficult to measure. The situation becomes more difficult for multi-lane motorways due to drivers lane-change behaviour. This research investigates statistical modelling and analysis of traffic flow. It contributes to the literature of transportation and traffic management and research in several aspects. First, it takes into account lane-changes in traffic modelling through incorporating a Markov chain model to describe the drivers lane-change behaviour. Secondly, the lane change probabilities between two adjacent lanes are not assumed to be fixed but rather they depend on the current traffic condition. A discrete choice model is used to capture drivers lane choice behaviour. The drivers choice probabilities are modelled by several traffic-condition related attributes such as vehicle time headway, traffic density and speed. This results in a highly nonlinear state equation for traffic density. To address the issue of high nonlinearity of the state space model, the EKF and UKF is used to estimate the traffic density recursively. In addition, a new transformation approach has been proposed to transform the observation equation from a nonlinear form to a linear one so that the potential approximation in the EKF & UKF can be avoided. Numerical studies have been conducted to investigate the performance of the developed method. The proposed method outperformed the existing methods for traffic density estimation in simulation studies. Furthermore, it is shown that the computational cost for updating the estimate of traffic densities for a multi-lane motorway is kept at a minimum so that online applications are feasible in practice. Consequently the traffic densities can be monitored and the relevant information can be fed into the traffic management system of interest.

Thesis (Master').

Σε αυτή τη διπλωματική εργασία, με τη χρήση ενός δείγματος που αποτελούνταν από 4458 καυκάσια άτομα, για τα οποία ήταν γνωστά: - ο Δείκτης Μάζας Σώματος (ΔΜΣ), - το φύλο, - η ηλικία, και - τα γενετικά χαρακτηριστικά 23 γονιδίων τους (δηλ. τα αλληλόμορφα γονίδια), διερευνήθηκε: - η ύπαρξη κύριων (μεμονωμένων) επιδράσεων αυτών των παραγόντων (φύλο, ηλικία, αλληλόμορφα γονίδια) στον ΔΜΣ, και - ο αντίκτυπος των αλληλεπιδράσεων μεταξύ έως και τριών παραγόντων στον ΔΜΣ. Μετά από την απαραίτητη προ-επεξεργασία των δεδομένων (αναζήτηση διπλών καταχωρήσεων και διαγραφή τους, κωδικοποίηση, εξαγωγή περιγραφικών στατιστικών στοιχείων, κ.λπ.) εκτελέστηκαν παραμετρικοί και μη παραμετρικοί στατιστικοί έλεγχοι με τη χρήση των 26 συνολικά διαθέσιμων μεταβλητών. Οι στατιστικές μέθοδοι που χρησιμοποιήθηκαν ήταν οι ακόλουθες (σε παρένθεση δίνονται κάποια συνοπτικά χαρακτηριστικά των μεταβλητών που ελέγχονταν κατά περίπτωση): - γραμμική συσχέτιση του Pearson (μεταξύ του ΔΜΣ και της Ηλικίας, καθώς και μεταξύ του ΔΜΣ και του Φύλου), - ανάλυση της διακύμανσης (ANOVA) (με χρήση του ΔΜΣ ως ποσοτική συνεχής μεταβλητή και των υπόλοιπων μεταβλητών σε κατηγορική μορφή), - μέθοδος x2 (chi-square): o με χρήση του ΔΜΣ σε κατηγορική μορφή 4 κατηγοριών (λιποβαρείς, κανονικού σωματικού βάρους, υπέρβαροι, παχύσαρκοι) και των υπόλοιπων μεταβλητών σε κατηγορική μορφή, και o με χρήση του ΔΜΣ σε κατηγορική μορφή 2 κατηγοριών [κανονικού βάρους (ενσωματώνει τις κατηγορίες λιποβαρών ατόμων και ατόμων κανονικού βάρους), μη κανονικού βάρους (ενσωματώνει τις κατηγορίες υπέρβαρων και παχύσαρκων ατόμων)] και των υπόλοιπων μεταβλητών σε κατηγορική μορφή, - μέθοδος μείωσης πολυπαραγοντικής διάστασης (MDR - Multifactor Dimensionality Reduction) με χρήση του ΔΜΣ σε κατηγορική μορφή 2 κατηγοριών [κανονικού βάρους (ενσωματώνει τις κατηγορίες λιποβαρών ατόμων και ατόμων κανονικού βάρους), μη κανονικού βάρους (ενσωματώνει τις κατηγορίες υπέρβαρων και παχύσαρκων ατόμων)] και των υπόλοιπων μεταβλητών σε κατηγορική μορφή. / In this thesis, using a sample that consisted of 4458 Caucasian men, for which we had the following available: - Body Mass Index (BMI) - Sex, - Age, and - The genetic characteristics of 23 genes (ie alleles) we examined: - The existence of major (individual) effects of these factors (gender, age, alleles) in BMI, and - The impact of interactions between up to three factors in BMI. After the necessary pre-processing of data (search for duplicate entries and deletion, coding, extraction of descriptive statistics, etc.) we performed parametric and non parametric statistical tests using the total of 26 available variables. The statistical methods used were as follows (in brackets are some brief features of the controlled variables as appropriate): - Linear correlation of Pearson (between BMI and age, and between BMI and sex) - Analysis of variance (ANOVA) (using BMI as a quantitative continuous variable and other variables in categorical form) - Chi-square test: o Using the BMI, in categorical form of four categories (underweight, normal weight, overweight, obese) and other variables in categorical form, and o Using the BMI, in categorical form of two categories [normal weight (incorporates underweight and normal weight), non-normal weight (incorporates overweight and obese)] and other variables in categorical form - Multifactor Dimensionality Reduction (MDR), using the BMI in categorical form of two categories [normal weight (incorporates underweight and normal weight), non-normal weight (incorporates overweight and obese)] and other variables in categorical form.