acase@tulane.edu

Binary outcomes are very common in medical and epidemiological studies. When scores are independent, logistic regression models are typically used. When outcomes are dependent, generalized estimating equations (GEE) methods are often used to analyze the correlated binary data. For any modeling procedure, an essential step is to determine goodness-of-fit (GoF) for the final model. For logistic regression, standard GoF tests have been developed and are available in most statistical software. The GoF statistics for the GEE method, however, have been developed more recently, and have not been incorporated into currently available software. Several GEE GoF statistics have been proposed (Barnhart & Williamson 1998; Horton et al. 1999; Pan 2002). The objective of this study was to compare these GEE GoF statistics using simulation data under different conditions. Sample sizes were varied, as were the possible covariates (discrete or continuous). Different models included time-dependent and time-independent covariates, quadratic components and interactions. Two or three scores per subject were generated with various correlations between scores No single GEE GoF statistic performed best under all conditions. All of them were poor at detecting the omission of the main effect for a binary time-dependent variable. Pan's statistics had the best performance in three cases: (1) detecting the omission of the interaction for binary time-dependent and time-independent covariates with two or three time-points; (2) detecting the omission of the interaction for a time-independent dichotomous variable and a time-dependent continuous variable with three time-points for exchangeable or auto-regressive correlated data; and (3) detecting the omission of a quadratic term in a time-dependent continuous variable with two time-points. Barnhart's statistics were also the most powerful statistics in the following three cases: (1) detecting the omission of the interaction for a time-independent dichotomous variable and a time-dependent continuous variable; (2) detecting the omission of the interaction for a time-independent dichotomous variable and a time-independent continuous variable; and (3) detecting the omission of the interaction for two continuous variables for two time-points. Horton's statistics had better performance in examining the omission of interaction for a time-independent dichotomous variable and a time-dependent continuous variable with three time-points independent data / acase@tulane.edu

acase@tulane.edu

In Part I, I have discussed the historical difficulties with understanding the Aristotelean concept of cause (aitia). I show that the Aristotelean concept of cause (aitia) is different from the 'usual' conception of causation, where the 'usual' conception of causation is that proposed by Hume or Reid. I also show how Aristotelean teleology gets wrongly associated with backward causation, design arguments, etc. In Part II, I show how contemporary functional explanations (in Larry Wright's sense of the term 'function') are really a subset of Aristotelean teleological explanations. In Part III, I argue that a living organism's particular genotype is the feature containing the potentials which are actualized through growth and change, making the adult form of the organism. In this way, I maintain that the genotype can be seen as the feature in which the efficient, formal, and final aitiai coincide, as Aristotle suggest they do in the Physics II.7. Thus, the genotype of an individual can be regarded as a teleological principle which directs the changes in the immature organism toward being a mature organism in accord with that organism's internal essence. This view, I argue, makes Aristotle compatible with contemporary genetic theory / acase@tulane.edu

acase@tulane.edu

acase@tulane.edu

acase@tulane.edu

acase@tulane.edu

acase@tulane.edu

The third-generation (3G) of mobile communications systems aim to provide enhanced voice, text and data services to the user. These demands give rise to the complexity and power consumption of the user equipment (UE) while the objective is smaller, lighter and power efficient mobiles. This thesis aims to examine ways of reducing the UE receiver’s computational cost while maintaining a good performance. One prominent multiple access scheme selected for 3G is code division multiple access. Receiver based multiuser detection techniques that utilise the knowledge of the downlink channel by the mobile have been extensively studied in the literature, in order to deal with multiple access and intersymbol interference. However, these techniques result in high mobile receiver complexity. Recently, work has been done on algorithms that transfer the complexity from the UE to the base station by exploiting the fact that in time division duplex mode the downlink channel can be known to the transmitter. By linear precoding of the transmitted signal the user equipment can be simplified to a filter matched to the user’s spreading code. In this thesis the problem of generic linear precoding is analysed theoretically and a method for analytical calculation of BER is developed. The most representative of the developed precoding techniques are described under a common framework, compared and classified as bitwise or blockwise. Bitwise demonstrate particular advantages in terms of complexity and implementation but lack in performance. Two novel bitwise algorithms are presented and analysed. They outperform significantly the existing ones, while maintain a reduced computational cost and realisation simplicity. The first, named inverse filters, is the Wiener solution of the problem after applying a minimum mean squared error criterion with power constraints. The second recruits multichannel adaptive algorithms to achieve the same goal. The base station emulates the actual system in a cell to converge iteratively to the pre-filters that precode the transmitted signals before transmission. The advantages and the performance of the proposed techniques, along with a variety of characteristics are demonstrated by means of Monte Carlo simulations.