Publicly available datasets in health science are often large and observational, in contrast to experimental datasets where a small number of data are collected in controlled experiments. Variables' causal relationships in the observational dataset are yet to be determined. However, there is a significant interest in health science to discover and analyze causal relationships from health data since identified causal relationships will greatly facilitate medical professionals to prevent diseases or to mitigate the negative effects of the disease. Recent advances in Computer Science, particularly in Bayesian networks, has initiated a renewed interest for causality research. Causal relationships can be possibly discovered through learning the network structures from data. However, the number of candidate graphs grows in a more than exponential rate with the increase of variables. Exact learning for obtaining the optimal structure is thus computationally infeasible in practice. As a result, heuristic approaches are imperative to alleviate the difficulty of computations. This research provides effective and efficient learning tools for local causal discoveries and novel methods of learning causal structures with a combination of background knowledge. Specifically in the direction of constraint based structural learning, polynomial-time algorithms for constructing causal structures are designed with first-order conditional independence. Algorithms of efficiently discovering non-causal factors are developed and proved. In addition, when the background knowledge is partially known, methods of graph decomposition are provided so as to reduce the number of conditioned variables. Experiments on both synthetic data and real epidemiological data indicate the provided methods are applicable to large-scale datasets and scalable for causal analysis in health data. Followed by the research methods and experiments, this dissertation gives thoughtful discussions on the reliability of causal discoveries computational health science research, complexity, and implications in health science research.

The fisher (Martes pennanti) is a medium sized member of the mustelid family that once roamed the forests of Washington and whose historic range in the western United States once spread throughout the northern Rocky Mountains, the Cascade and Coast Ranges, and the Sierra Nevada (Carroll, Zielinski, and Noss 1999; Powell 1993, Spencer et al. 2011). Due to pressures from trapping and habitat fragmentation, the abundance of the species in the western United States has decreased dramatically and is thought to be limited to several small, isolated populations. In 2008, fishers were reintroduced to the Olympic Peninsula; however, bobcat (Lynx rufus) predation in the first years is thought to have killed off a significant portion of the released fisher hindering their ability to establish a self-sustaining population (Lewis et al. 2011). Other studies in the western United States have shown that bobcats can be a dramatic force on small or isolated fisher populations. The coniferous forest of the southern Washington Cascades is the possible site of a release of currently extirpated fishers. My research examines the distribution of bobcats in the region and explores the implication this and the habitat variables of the area have for a future reintroduction of fisher. The workflow of the research was a stepwise process of: 1) surveying forested areas in the southern Washington Cascades for the presence and absence of bobcat and acquiring previously completed survey data 2) using a classification tree to model the correlation of bobcat presence or absence with forest variables and 3) applying these relationships to spatial analysis the creation of maps showing areas of high ranking fisher habitat. The classification tree modeled the correlation between the forest variables and the results of the surveys, which included 145 bobcat absence observations and 39 presence observations. The model highlighted a 95% probability of absence above 1,303 m in elevation, 73% probability of absence in areas under 1,303 m in elevation and with a tree diameter value under 43.45 cm, 57% probability of absence in areas between 1,070 m and 1,303 m in elevation and with a tree diameter value above 43.45 cm, and an 89% probability of bobcat presence in areas under 1,070 m in elevation with a tree diameter value above 43.45 cm. I applied an upper elevation limit of 1,676 meters as a threshold for suitable habitat and only considered habitat suitable in cells with a tree diameter above 29 cm. The three locations highlighted as the most suitable areas for reintroduction due to a large amount of the highest ranking habitat and the largest aggregations of suitable habitat cells were around the William O. Douglas Wilderness that straddles the border of the Gifford Pinchot National Forest (GPNF) and the Wenatchee National Forest, another location in the Norse Peak Wilderness northeast of Mount Rainier, and a third location in Indian Heaven Wilderness in the southern portion of the GPNF.

Internet technology has affected humans' lives in many disciplines. The search engine is one of the most important Internet tools in that it allows people to search for what they want. Search queries entered in a web search engine can be used to predict dengue incidence. This vector borne disease causes severe illness and kills a large number of people every year. This dissertation utilizes the capabilities of search queries related to dengue and climate to forecast the number of dengue cases. Several machine learning techniques are applied for data analysis, including Multiple Linear Regression, Artificial Neural Networks, and the Seasonal Autoregressive Integrated Moving Average. Predictive models produced from these machine learning methods are measured for their performance to find which technique generates the best model for dengue prediction. The results of experiments presented in this dissertation indicate that search query data related to dengue and climate can be used to forecast the number of dengue cases. The performance measurement of predictive models shows that Artificial Neural Networks outperform the others. These results will help public health officials in planning to deal with the outbreaks.

We present an innovative Bayesian method for estimation of statistical parameters of time series data. This method works by comparing coefficients of Karhunen-Lo\`{e}ve expansion of observed and synthetic data with known parameters. We show one new method for generating synthetic data with prescribed properties and we demonstrate on a numerical example how this method can be used for estimation of physically interesting features in power spectra calculated from observed light curves of some X-ray sources.

The statistical assessment of harvested fish populations, such as the Atlantic-Iberian sardine (AIS) stock, needs to deal with uncertainties inherent in fisheries systems. Uncertainties arising from sampling errors and stochasticity in stock dynamics must be incorporated in stock assessment models so that management decisions are based on realistic evaluation of the uncertainty about the status of the stock. The main goal of this study is to develop a stock assessment framework that accounts for some of the uncertainties associated with the AIS stock that are currently not integrated into stock assessment models. In particular, it focuses on accounting for the uncertainty arising from the catch data sampling process. The central innovation the thesis is the development of a Bayesian integrated stock assessment (ISA) model, in which an observation model explicitly links stock dynamics parameters with statistical models for the various types of data observed from catches of the AIS stock. This allows for systematic and statistically consistent propagation of the uncertainty inherent in the catch sampling process across the whole stock assessment model, through to estimates of biomass and stock parameters. The method is tested by simulations and found to provide reliable and accurate estimates of stock parameters and associated uncertainty, while also outperforming existing designed-based and model-based estimation approaches. The method is computationally very demanding and this is an obstacle to its adoption by fisheries bodies. Once this obstacle is overcame, the ISA modelling framework developed and presented in this thesis could provide an important contribution to the improvement in the evaluation of uncertainty in fisheries stock assessments, not only of the AIS stock, but of any other fish stock with similar data and dynamics structure. Furthermore, the models developed in this study establish a solid conceptual platform to allow future development of more complex models of fish population dynamics.

No description available.

My dissertation consists of three chapters. In the first chapter, I analyze theeffect of the merger between American Airlines (AA) & US Airways (US) on market price and product quality. I use two complementary methodologies: difference-in-differences (DID) and merger simulation. Contrary to other results in the airline literature, the DID analysis shows that, overall, price has decreased as a result of themerger. While divestitures required as part of the merger had a strong price-reducing effect, the overall decrease involves non-divestiture markets as well. Interestingly, the decrease appears only in large airport-pair markets, whereas prices rose considerably in smaller ones. Effects on quality are mixed. The DID analysis shows that the merger reduced flight cancellations, increased flight delays, and had no effect on flight frequency or capacity overall. Using merger simulation, I find that the change in ownership leads to a 3% increase in price. The structural model performs betterin predicting the post-merger price if I allow the model to deviate from the Bertrand-Nash conduct. A 10% cost reduction due to the merger is able to predict the post-merger price quite well. When I incorporate a conduct parameter into the model, the required percentage of cost savings is lower. Given the divestiture and the subsequententry of low-cost carriers (LCCs), tacit collusion may break down. Thus both cost savings and reduced cooperation could explain a reduction in the price in the post-merger period.<div><br></div><div>In my second chapter, I analyze possible reasons why airline prices are higher inthe smaller markets compared to larger markets. In the literature, most of the studies ignore the fact that the smaller markets are different compared to larger markets in terms of the nature of competition. I find that a combination of lower competition, and lack of entry from low cost carriers (LCCs) are the reasons behind higher prices in the smaller city-pair markets. I show that price is substantially higher in a market with a fewer number of firms controlling for several other factors. My paper estimates the modified critical number of firms to be 5 and the critical value of the HHI to be .6.<br><div><br></div><div>In my third chapter, I study the effect of announcement of investment in research & development (R&D) on the value of a firm in the pharmaceutical industry. Three types of R&D by the pharmaceutical firms are considered for the analysis: acquisition of other smaller firms, internal investment in R&D, and collaborative investment in R&D. This chapter finds that few target specific characteristics and financial charac-teristics of the acquiring firm are important drivers of the abnormal returns around the announcement period.<br></div><div><br></div></div>

A dissertation submitted to the Faculty of Science, University of the Witwatersrand, in fulfilment of the requirements for the degree of Master of Science. Johannesburg, 2017. / With an increase in opportunistically-collected data, statistical methods that can accommodate unstructured designs are increasingly useful. Spatial capturerecapture (SCR) has such potential, but its applicability for species that are strongly gregarious is uncertain. It assumes that average animal locations are spatially random and independent, which is violated for gregarious species. I used a data set for African elephants (Loxodonta africana) and data simulation to assess bias and precision of SCR population density estimates given violations in location independence. I found that estimates were negatively biased and likely too precise if non-independence was ignored. Encounter heterogeneity models produced more realistic precision but density estimates were positively biased. Lowest bias was achieved by estimating density of groups, group size, and then multiplying to estimate overall population density. Such findings have important implications for the reliability of population density estimates where data are collected by unstructured means. / LG2017

A research report submitted to the Faculty of Engineering and the Built Environment, University of the Witwatersrand, Johannesburg, in partial fulfilment of the requirements for the degree of Master of Science in Engineering. Date: September 2017, Johannesburg / The focus in this research investigation is to investigate the Process Parameter Optimisation in Laser Beam Forming (LBF) process using the 4.4 kW Nd: YAG laser system – Rofin DY 044 to form 200 x 50 x 3 mm3 mild steel - AISI 1008 samples. The laser power P, beam diameter B, scan velocity V, number of scans N, and cooling flow C were the five input parameters of interest in the investigation because of their influence in the final formed product. Taguchi Design of Experiment (DoE) was used for the selection and combination of input parameters for LBF process. The investigation was done experimentally and computationally. Laser Beam Forming (LBF) input parameters were categorised to three different levels, low (L), medium (M), and high (H) laser forming (LBF) parameters to evaluate parameters that yield maximum bending and better surface finish/quality. The conclusion drawn from LBF process is that samples which are LBFormed using low parameter settings had unnoticeable bending and good material surface finishing. On the other hand, samples LBFormed using medium parameters yielded visible bending and non-smooth surface finishing, while samples processed using high LBF parameters yielded maximum bending and more surface roughness than the other two process parameters. / MT2018

A research report submitted to the Faculty of Science, in partial fulfilment of the requirements for the degree of Master of Science, University of the Witwatersrand, Johannesburg. 30 October 2014. / The reporting of non-financial data has steadily increased over the past three decades and there is evidence that including social and environmental indicators in the annual report is correlated with improved environmental performance of listed companies. The annual reports of a selection of 82 JSE-listed companies, including the full JSE Top 40, were analysed for mentions of the natural environment for the reporting periods of 2010 and 2012. The introduction of the King III principles by the JSE occurred between these two periods, providing an opportunity to assess the impacts that this move had on annual reporting. Attention was paid to mentions in the leadership reviews by the Chairmen and the CEOs, presence of empirical environmental data, environmental KPIs and the manner in which these data were presented and discussed in the report. In addition, a survey asking qualitative details of company reporting policy was conducted among the staff members responsible for environmental reporting of these companies. The standard and sophistication of environmental reporting varied widely across the sample, with Top 40 companies generally reporting better than non-Top 40 companies. Primary industries were more likely to provide empirical data than service industries and only agricultural industries appeared concerned with the manner in which changes in the natural environment could affect their business. There remains a wide variation in the type and detail of environmental reporting across the sample with very little evidence that the data, as reported, play a meaningful role in the decisions of either management or investors.