<p>The accumulation of high-throughput data from vast sources has drawn a lot attentions to develop methods for extracting meaningful information out of the massive data. More interesting questions arise from how to combine the disparate information, which goes beyond modeling sparsity and dimension reduction. This dissertation focuses on the innovations in the area of heterogeneous data integration.</p><p>Chapter 1 contextualizes this dissertation by introducing different aspects of meta-analysis and model frameworks for high-dimensional genomic data.</p><p>Chapter 2 introduces a novel technique, joint Bayesian sparse factor analysis model, to vertically integrate multi-dimensional genomic data from different platforms. </p><p>Chapter 3 extends the above model to a nonparametric Bayes formula. It directly infers number of factors from a model-based approach.</p><p>On the other hand, chapter 4 deals with horizontal integration of diverse gene expression data; the model infers pathway activities across various experimental conditions. </p><p>All the methods mentioned above are demonstrated in both simulation studies and real data applications in chapters 2-4.</p><p>Finally, chapter 5 summarizes the dissertation and discusses future directions.</p> / Dissertation

Electronic Passive Component ¡]E.P.C.¡^industry has been developed for over 40 years in local industries. Which created the raise of the early Taiwan¡¦s Electronic Component Industry. In the recent few years, Taiwan¡¦s E.P.C. industry has rapidly grown as well as 3C Industries generation by turns from worldwide. Although the market share of Taiwan¡¦s E.P.C. industry seems growing as well, but the serious weakness of the low entry barrier of the low end products, the competition of cost price reduction, lacking of the capability of core materials and Research ¡®Development and the consolidated role in Mainland China will deeply threaten the exists of Taiwan E.P.C.. Based on case studies from E.P.C. industries and domestic firms. The thesis explored the success experience of competitive advantage from the case research of utilizing The Six Force Factor Analysis model and The SWOT Analysis model and also adopted several competitive strategies to recommend. The consequence were concluded as follows¡G 1.To the Government¡Gto promote and cooperate the local E.P.C. industry to obtain the advantage of global resource of high technical staff and abroad market. 2.To Taiwan E.P.C. industry¡Gto make a advantaged alliance to enhance the marketing strategies of one time shop and solve the weakness of R¡®D. 3.To the research firm¡Gto fulfill the value chain advantage in forging an alliance with the core materials resource and target market oriented, and flexible adjusting the strategy of production lines cooperation between Mainland China and Taiwan.

In developing countries, there is a great demand for university education. In order to select students to universities a standardized test score is used. In Turkey, the Student Selection Test (SST) have important role in admission to universities. However, there is very limited knowledge about what SST mathematics sections actually measures. The main purpose of the present study is to evaluate the content of the mathematics subtest of the SST in line with mathematical cognitive skills and eventually provide construct related evidence for dimensionality of the test items. Within this framework, it is aimed to cross validate the mathematics subtest across gender groups, school types and two consecutive years. Also relations among mathematical abilities are investigated. This study is first in investigating what is measured by SST Mathematics sections and analyzing construct validity by testing several nested confirmatory factor models. Comparison of fit indices of five competitive models showed three-factor model has better fit indices in which Basic Computation Ability, Advanced Computation Ability and Geometry Ability is measured. It is concluded that problem solving items are not measuring a different process, but measures some sort of computation ability. There is a problem related to the content of the mathematics subtests of the SST in line with mathematical cognitive skills. Higher order cognitive skills are not measured properly. Three-factor model is tested about the invariance of the factors across gender, school types and years. It is concluded that invariant factor structure indicates that SST mathematics section is operating similarly for subgroups and years. The relations among mathematical abilities on three-factor model are investigated by item mapping and structural equation models. It is seen that Basic Computation Ability is a prerequisite to acquire Geometry Ability and Advanced Computation Ability.

In the present study, we compare factor analytic models of post-traumatic stress disorder (PTSD) in terms of their fit and predictive utility with regard to external correlates such as comorbid diagnoses and other psychosocial outcomes. Competing models were compared and validated in an epidemiological dataset (N = 23,936). Confirmatory factor analyses (CFA) using models from prior literature with four through seven factor solutions were conducted. The seven factor Hybrid model, the six-factor Anhedonia model, and the six-factor Externalizing Behaviors model were the first, second, and third best-fitting models, respectively; however, the inconsistency of associations with external correlates and high factor intercorrelations suggested that higher-factor solutions may sacrifice parsimony for minimal gains in utility. The Anhedonia and Hybrid models’ separate Anhedonia and Negative Affect factors (a core difference from other models) demonstrated limited utility in differentially associating with distinct constructs under the internalizing umbrella. Anhedonia and Negative Affect also correlated highly with each other and nearly perfectly with the factors composed of their combined symptoms (e.g. the Externalizing Behaviors model’s Numbing factor), suggesting a "lumped" factor would be more parsimonious. The Externalizing Behaviors model showed predictive utility in accounting for externalizing comorbidities as well as differentiating among constructs within the internalizing spectrum; however, it lacked robust associations with externalizing behavioral outcomes such as frequency and quantity of drinking. These results give reason for concern that predominant structural models of PTSD may not be adequate for discriminating among or predicting functional outcomes related to PTSD symptomatology in trauma-exposed populations.<br>