A composite likelihood consists of a combination of valid likelihood objects, and in particular it is of typical interest to adopt lower dimensional marginal likelihoods. Composite marginal likelihood appears to be an attractive alternative for modeling complex data, and has received increasing attention in handling high dimensional data sets when the joint distribution is computationally difficult to evaluate, or intractable due to complex structure of dependence. We present some aspects of methodological development in composite likelihood inference. The resulting estimator enjoys desirable asymptotic properties such as consistency and asymptotic normality. Composite likelihood based test statistics and their asymptotic distributions are summarized. Higher order asymptotic properties of the signed composite likelihood root statistic are explored. Moreover, we aim to compare accuracy and efficiency of composite likelihood estimation relative to estimation based on ordinary likelihood. Analytical and simulation results are presented for different models, which include multivariate normal distributions, times series model, and correlated binary data.

No description available.

The overall goal of the present research is to provide a computationally based methodology to realize the projected extraordinary properties of Carbon Nanotube (CNT)- reinforced composites and polymeric nanofibers for engineering applications. The discovery of carbon nanotubes (CNT) and its derivatives has led to considerable study both experimentally and computationally as carbon based materials are ideally suited for molecular level building blocks for nanoscale systems. Research in nanomechanics is currently focused on the utilization of CNTs as reinforcements in polymer matrices as CNTs have a very high modulus and are extremely light weight. The nanometer dimension of a CNT and its interaction with a polymer chain requires a study involving the coupling of the length scales. This length scale coupling requires analysis in the molecular and higher order levels. The atomistic interactions of the nanotube are studied using molecular dynamic simulations. The elastic properties of neat nanotube as well as doped nanotube are estimated first. The stability of the nanotube under various conditions is also dealt with in this dissertation. The changes in the elastic stiffness of a nanotube when it is embedded in a composite system are also considered. This type of a study is very unique as it gives information on the effect of surrounding materials on the core nanotube. Various configurations of nanotubes and nanocomposites are analyzed in this dissertation. Polymeric nanofibers are an important component in tissue engineering; however, these nanofibers are found to have a complex internal structure. A computational strategy is developed for the first time in this work, where a combined multiscale approach for the estimation of the elastic properties of nanofibers was carried out. This was achieved by using information from the molecular simulations, micromechanical analysis, and subsequently the continuum chain model, which was developed for rope systems. The continuum chain model is modified using properties of the constituent materials in the mesoscale. The results are found to show excellent correlation with experimental measurements. Finally, the entire atomistic to mesoscale analysis was coupled into the macroscale by mathematical homogenization techniques. Two-scale mathematical homogenization, called asymptotic expansion homogenization (AEH), was used for the estimation of the overall effective properties of the systems being analyzed. This work is unique for the formulation of spectral/hp based higher-order finite element methods with AEH. Various nanocomposite and nanofibrous structures are analyzed using this formulation. In summary, in this dissertation the mechanical characteristics of nanotube based composite systems and polymeric nanofibrous systems are analyzed by a seamless integration of processes at different scales.

No description available.

In this work, we propose, implement and test two fully automated mesh adaptation methods for 1-D multigroup eigenproblems. The first method is the standard hp-adaptive refinement strategy and the second technique is a goal-oriented hp-adaptive refinement strategy. The hp-strategies deliver optimal guaranteed solutions obtained with exponential convergence rates with respect to the number of unknowns. The goal-oriented method combines the standard hp-adaptation technique with a goal-oriented adaptivity based on the simultaneous solution of an adjoint problem in order to compute quantities of interest, such as reaction rates in a sub-domain or point-wise fluxes or currents. These algorithms are tested for various multigroup 1-D diffusion problems and the numerical results confirm the optimal, exponential convergence rates predicted theoretically.

The Reproducing Kernel Element Method (RKEM) implementation of the Fleck-Hutchinson phenomenological strain gradient theory in 1D, 2D and 3D is implemented in this research. Fleck-Hutchinson theory fits within the framework of Touplin- Mindlin theories and deals with first order strain gradients and associated work conjugate higher-order stress. Theories at the intrinsic or material length scales find applications in size dependent phenomena. In elasticity, length scale enters the constitutive equation through the elastic strain energy function which depends on both strain as well as the gradient of rotation and stress. The displacement formulation of the Touplin Mindlin theory involve diffrential equations of the fourth order, in conventional finite element method C1 elements are required to solve such equations, however C1 elements are cumbersome in 2D and unknown in 3D. The high computational cost and large number of degrees of freedom soon place such a formulation beyond the realm of practicality. Recently, some mixed and hybrid formulations have developed which require only C0 continuity but none of these elements solve complicated geometry problems in 2D and there is no problem yet solved in 3D. The large number of degrees of freedom is still inevitable for these formulation. As has been demonstrated earlier RKEM has the potential to solve higher-order problems, due to its global smoothness and interpolation properties. This method has the promise to solve important problems formulated with higher order derivatives, such as the strain gradient theory.

We study three category-theoretic types in the context of functional query languages (typed lambda-calculi extended with additional operations for bulk data processing). The types we study are: / Engineering and Applied Sciences

The philosophical and scientific studies of consciousness are two disciplines having much to learn from one another. On the one hand, a science of consciousness involves taking an objective approach to what is essentially a subjective phenomenon, giving rise to tricky conceptual and methodological issues, ones an analytic philosopher is perhaps best equipped to handle. On the other hand, a wealth of data now exists concerning the neural basis of consciousness. Such data, interpreted properly, can confirm or disconfirm philosophical views on consciousness, helping adjudicate debates thus far intractable. This dissertation explores some ways in which the philosophy and science of consciousness can be of mutual benefit to one another.

This work presents a robust algorithm for voice activity detection (VAD) and noise reduction mechanism using combined properties of higher-order statistics (HOS) and an efficient algorithm to estimate the instantaneous Signal-to-Noise Ratio (SNR) of speech signal in a background of acoustic noise. The flat spectral feature of Linear Prediction Coding (LPC) residual results in distinct characteristics for the cumulants in terms of phase, periodicity and harmonic content and yields closed-form expressions for the skewness and kurtosis. The HOS of speech is immune to Gaussian noise and this makes them particularly useful in algorithms designed for low SNR environments. The proposed algorithm uses HOS and smooth power estimate metrics with second-order measures, such as SNR and LPC prediction error, to identify speech and noise frames. A voicing condition for speech frames is derived based on the relation between the skewness, kurtosis of voiced speech and estimate of smooth noise power. The algorithm presented and its performance is compared to HOS-only based VAD algorithm. The results show that the proposed algorithm has an overall better performance, with noticeable improvement in Gaussian-like noises, such as street and garage, and high to low SNR, especially for probability of correctly detecting speech. The proposed algorithm is replicated on DSK C6713.

No description available.