No description available.

This thesis is concerned with the problem of how people learn to use uncer­tain information for making judgments. The general framework for the thesis is Social Judgment Theory (SJT). First the S3T paradigm, and some research conducted within the paradigm, is briefly described, and a series of four empirical studies is summarized. The studies are concerned with two factors that have been found to have great effect on subjects achievement in cue probability learning (CPL) tasks: task predictability, and the form of the function relating cue and criterion. The effects of these two factors were studied in experiments employing cue-probability learning tasks. The studies concerned with task predictability addressed the following questions (a) Do subjects understand the probabilistic nature of CPL-tasks? (b) Are subjects able to detect that a random task is, in fact, random, a study undertaken to test an aspect of Seligmans "theory of helplessness". This was also an attempt to bring emotional factors more in foeus.(c) Do subjects use data from the task only to test hypotheses, or do they use data also to construct hypotheses? The results showed that (a) subjects do not seem to be able to cope with probabilistic tasks in an optimal statistical manner. Instead they seem to use a deterministic approach to the tasks, because they do not understand the probabilistic nature of the task, (b) Task predictability affecs subjects mood, but not in the way predicted by Seligman, (c) Subjects seem to use data frorn the task only to test their hypotheses. The results thus supported the hypo­theses sampling model for the learning of CPL-tasks. As for the factor of function form, the following questions were addressed, (a) What hypotheses about relations between variables do subjects have? (b) Is the difficulties subjects have in learning complex rules in CPL-tasks due to a low availability of hypotheses about complex rules? The results showed that, (a) the hypothesis hierarchy as revealed in the present experiments was in general agreement with earlier results. However, few nonlinear hypotheses were observed, and other rules than functional rules were observed, (b) The difficulties subjects have to learn complex rules in CPL-tasks do not seem to be caused by low availability of rules. Finally, some suggestions are given for how the SJT-paradigm should be deve­loped. Specifically, it is suggested that the effects of emotional factors should be given more attention, and that the paradigm should be turned into a more general hypothesis testing model / <p></p><p>　</p><p></p> / digitalisering@umu

This thesis investigates extending the dynamically-typed, functional programming language Scheme, with simulated overloading in order to permit the binding of multiple, distributed defnitions to function names. Overloading facilitates the use of an incremental style of programming in which functions can be defined with a base behaviour and then extended with additional behaviour as it becomes necessary to support new data types. A technique is demonstrated that allows existing functions to be extended, without modifcation, therefore improving code reuse. Using the primitives provided by Scheme, it is possible to write functions that perform like the generic routines (functions) of the programming language EL1. These functions use the type of their arguments to determine, at run-time, the computation to perform. It is shown that by gathering the definitions for an overloaded function and building a generic routine, the language appears to provide overloading. A language extension that adds the syntax necessary to instruct the system to gather the distributed set of definitions for an overloaded function and incrementally build an equivalently applicable generic function is described. A simple type inference algorithm, necessary to support the construction of generic functions, is presented and detailed. Type inference is required to determine the domain of an overloaded function in order to generate the code needed to perform run-time overload resolution. Some limitations and possible extensions of the algorithm are discussed.

The view that knowledge-yielding single-premise deductive inference must proceed from a known premise is very plausible at first blush. In this thesis I explore in detail the possibility that this view is false. I construct a series of challenging cases against the principle of Counter-Closure, which expresses this view. These cases force theorists endorsing a variety of contemporary views to either (i) abandon Counter-Closure; (ii) admit into their epistemology novel and theory-specific kinds of Gettier cases; or (iii) make significant revisions to their theories. I offer considerations that help would-be deniers of Counter-Closure explain away its prima facie plausibility and suggest a suitable theoretical replacement phrased in terms of justification rather than knowledge. Finally, I connect this discussion with debates in the epistemologies of testimony and memory, where analogue principles to Counter- Closure have been recently subjected to critical scrutiny.

Network structure plays a significant role in determining the performance of network inference tasks. An interactive tool to study the dependence of network topology on estimation performance was developed. The tool allows end-users to easily create and modify network structures and observe the performance of pole estimation measured by Cramer-Rao bounds. The tool also automatically suggests the best measurement locations to maximize estimation performance, and thus finds its broad applications on the optimal design of data collection experiments. Finally, a series of theoretical results that explicitly connect subsets of network structures with inference performance are obtained.

Leung Shun-piu. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2001. / Includes bibliographical references (leaves 41-43). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Multiplicity Problem in Multiple Testing --- p.1 / Chapter 1.2 --- Family --- p.2 / Chapter 1.3 --- Family wise Error Rate --- p.2 / Chapter 1.4 --- Multiple Comparisons with a Control --- p.3 / Chapter 1.5 --- Single-step Procedures vs Stepwise Procedures --- p.4 / Chapter 1.6 --- Dunnett Procedure --- p.5 / Chapter 1.6.1 --- One-way Fixed Effect Model --- p.5 / Chapter 1.6.2 --- Simultaneous Inference and Test Statistics --- p.6 / Chapter 1.6.3 --- Calculation of the Upper and Lower Percentage Points --- p.8 / Chapter 1.7 --- Objectives --- p.10 / Chapter 2 --- Testing Procedures --- p.11 / Chapter 2.1 --- Simultaneous Inference in Mixed Families --- p.11 / Chapter 2.2 --- "Evaluation of C1,α and c2,α" --- p.13 / Chapter 2.3 --- Extension to Mixed Families with 3 Testing Groups --- p.15 / Chapter 3 --- The Calculation of Critical Values --- p.18 / Chapter 3.1 --- Calculation of Critical Values --- p.18 / Chapter 3.2 --- "Tabulation of Critical Values (c-*,α c*,α)" --- p.22 / Chapter 4 --- Numerical Example --- p.30 / Chapter 5 --- Conclusions --- p.34 / Appendix --- p.35 / References --- p.41

A reliable inference of networks from data is of key interest in many scientific fields. Several methods have been suggested in the literature to reliably determine links in a network. These techniques rely on statistical methods, typically controlling the number of false positive links, but not considering false negative links. In this thesis new methodologies to improve network inference are suggested. Initial analyses demonstrate the impact of false positive and false negative conclusions about the presence or absence of links on the resulting inferred network. Consequently, revealing the importance of making well-considered choices leads to suggest new approaches to enhance existing network reconstruction methods. A simulation study, presented in Chapter 3, shows that different values to balance false positive and false negative conclusions about links should be used in order to reliably estimate network characteristics. The existence of type I and type II errors in the reconstructed network, also called biased network, is accepted. Consequently, an analytic method that describes the influence of these two errors on the network structure is explored. As a result of this analysis, an analytic formula of the density of the biased vertex degree distribution is found (Chapter 4). In the inverse problem, the vertex degree distribution of the true underlying network is analytically reconstructed, assuming the probabilities of type I and type II errors. Chapters 4-5 show that the method is robust to incorrect estimates of α and β within reasonable limits. In Chapter 6, an iterative procedure to enhance this method is presented in the case of large errors on the estimates of α and β. The investigations presented so far focus on the influence of false positive and false negative links on the network characteristics. In Chapter 7, the analysis is reversed - the study focuses on the influence of network characteristics on the probability of type I and type II errors, in the case of networks of coupled oscillators. The probabilities of α and β are influenced by the shortest path length and the detour degree, respectively. These results have been used to improve the network reconstruction, when the true underlying network is not known a priori, introducing a novel and advanced concept of threshold.

Pendant mes travaux de thèse, j'ai développé et appliqué des méthodes informatiques utilisant des données de réseaux afin d'aider l'analyse des données biologiques à haut-débit. Ma thèse consiste en trois projets : L'identification de protéines supplémentaires dans des approches de protéomique différentielle à l'aide des réseaux d'interaction protéiques, l'identification de réseaux régulatoires sous-jacents aux réponses aux stress abiotiques dans arabidopsis thaliana et l'analyse de signature transcriptomique de réponse immunitaire d'hôte spécifique à différentes étapes d'infection par shigella flexneri. / In the course of my Ph.D work, i have developed and applied methods making use of network information to adavance the analysis of high-throughput biological data. My thesis comprises three projects :- The identification of additional proteins in differential protemics using protein interaction networks. In this study, we developed a novel computational approach based on protein-protein interaction networks to identify a list of proteins that might have remained undetected in differential proteomic profiling experiments.- The transcriptional regulatory networks underlying responses to environmental stresses. Based on publicly available data, measuring the response of A. Thaliana to a set of abiotic stresses in a time-resolved manner, we applied two complimentary approaches to derive gene regulatory networks underlying the plant's response to the perceived stresses.- The analysis of transcriptional host immune response signatures specific for distinct stages of infection by shigella flexneri. During their host invasion process, shigella localize to different subcellular niches.

The thesis deals with Bayesian approach to statistics and its comparison to frequentist approach. The main aim of the thesis is to compare frequentist and Bayesian approaches to statistics by analyzing statistical inferences, examining the question of subjectivity and objectivity in statistics. Another goal of the thesis is to draw attention to the importance and necessity to teach Bayesian statistics at our University more profound. The thesis includes three chapters. The first chapter presents a Bayesian approach to statistics and its main notions and principles. Statistical inferences are treated in the second chapter. The third chapter deals with comparing Bayesian and frequentist approaches. The final chapter concerns the place of Bayesian approach nowadays in science. Appendix concludes the list of Bayesian textbooks and Bayesian free software.

La reconstruction des réseaux de régulation génétique (GRNs) est une étape importante pour la compréhension des mécanismes de régulation complexes régissant le fonctionnement de la cellule. De nombreuses approches de modélisation ont été introduites pour inférer le lien de causalité entre les gênes à l'aide des données d'expression génétiques. Cependant, les performances de ces approches sont limitées principalement a causé a des données de grande dimension. En plus, ces méthodes ne restent pas généralement la réalité biologique ou l'interaction entre gêne est réalisée d'une manière coopérative mais considèrent un modèle plus simple ou seules les interactions binaires sont considérées. Dans cette thèse, nous présentons de nouvelles méthodes d'inférence de GRN coopératifs afin améliorer la stabilité et la précision de la reconstruction des GRN en utilisant des techniques d'ensemble. Pour un gêne cible donne, nous extrayons un ensemble de GRNs coopératifs a partir de données discrétisées d'expression. Les GRNs ainsi que les interactions génétiques inférés sont classés selon leur importance en utilisant la régression linéaire sur la base des données d'expression continues. Les évaluations menées sur les données du challenge Dream5 et sur des données humaines de cancer de la vessie démontrent que nos méthodes sont efficaces, tout particulièrement si la taille des données d'apprentissage est petite. / Reconstruction of Gene Regulatory Networks (GRNs) is an important step towards understanding the complex regulatory mechanisms within the cell. Many modeling approaches have been introduced ti find the causal relationship between genes using expression data. However, they surfer from the high dimensionality problem i.e., having a large number of genes but a small number of samples negatively impacts the results. Moreover, these models do not truthfully reflect the biological system where genes interactions are performed in a cooperative manner but rather simplify the problem by tackling only binary interactions. In this thesis, we present new methods for cooperative GRN inference to improve the stability and accuracy of GRNs reconstruction leveraging ensemble methods. For a given target gene, we extract an ensemble of GRNs from discretized expression data. Inferred networks are then evaluated by ranking individual regulation relationships using a regression based technique and continuous expression data. Evaluations on Dream5 challenge data as well as human cancer data demonstrate that our methods are effcient, especially when operating on a small data set.