No description available.

No description available.

Construct truncation can be defined as the failure to capture variation along the entire continuum of a construct reliably. It can occur due to suboptimal person selection or due to suboptimal item selection. In this thesis, I used a series of simulation studies coupled with real data examples to characterise the consequences of construct truncation on the inferences made in empirical research. The analyses suggested that construct truncation has the potential to result in significant distortions of substantive conclusions. Based on these analyses I developed recommendations for anticipating the circumstances under which construct truncation is likely to be problematic, identifying it when it occurs, and mitigating its adverse effects on substantive conclusions drawn from affected data.

Ma, Chung Ho. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2010. / Includes bibliographical references (leaves 56-57). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Additive model --- p.3 / Chapter 1.2 --- Multiplicative model --- p.4 / Chapter 2 --- Review of benchmarking methods --- p.7 / Chapter 2.1 --- Regression Method and Signal Extraction Method (Additive Case) --- p.7 / Chapter 2.1.1 --- Regression method for Additive model --- p.8 / Chapter 2.1.2 --- Signal extraction method with known auto-covariance of signal for Additive model --- p.9 / Chapter 2.1.3 --- Signal extraction method with unknown auto-covariance of signal for Additive model --- p.10 / Chapter 2.2 --- Regression Method and Signal Extraction Method (Multiplicative Case) --- p.15 / Chapter 2.2.1 --- Regression method for Multiplicative model --- p.15 / Chapter 2.2.2 --- Signal extraction method for Multiplicative model --- p.17 / Chapter 3 --- Benchmarking using an assigned error model --- p.21 / Chapter 3.1 --- Model setting --- p.22 / Chapter 3.2 --- Simulation procedures --- p.24 / Chapter 3.3 --- Simulation results --- p.26 / Chapter 3.3.1 --- Comparison of benchmarking methods for multiplicative model --- p.34 / Chapter 3.3.2 --- Suggestions on the selection of an assigned error model --- p.39 / Chapter 3.4 --- Summary on the effects of CV and Φ --- p.44 / Chapter 4 --- Error modelling by using benchmarks --- p.47 / Chapter 4.1 --- Benchmark Forecasting Method --- p.47 / Chapter 4.2 --- Benchmarking using an estimated error model --- p.49 / Chapter 4.3 --- Comparison of the best Φ and the estimated Φ on the benchmarking prediction --- p.53 / Chapter 5 --- Conclusion --- p.54 / Bibliography --- p.56

This thesis introduces a new algorithm to search for optimal spatial sampling design. It is found in previous studies of Zhu and Stein (2006) that the optimal sampling design for spatial prediction with estimated parameters is nearly regular with a few clustering points. The pattern is similar to the generalized Neyman-Scott (GNS) process introduced by Yau and Loh (2012), which allows for regularity in the parent process. This motivates the use of a realization of the GNS process as a spatial sampling design. This method translates the high dimensional optimization problem of selecting sampling sites into a low dimensional optimization problem of searching for the optimal parameter sets in the GNS process. Simulation studies indicate that the proposed sampling design algorithm is more computationally efficient while the result of criterion minimization is comparable to traditional methods. / 本文介紹了一種新的算法來搜索最優空間採樣設計。先前Zhu和Stein（2006）的研究發現，按被估計參數的空間預測的最優採樣設計是近乎有規律的，同時伴隋一些聚類點。該圖案與Yau和Loh（2012）介紹的廣義Neyman-Scott（GNS）過程相似，其中的父過程擁有規律性。這驅使我們使用GNS過程的實現作為空間採樣設計。這種方法把選擇採樣點的高維優化問題轉化為搜索最優GNS過程參數集的低維優化問題。模擬實驗顯示，該採樣設計的算法是計算效率更高，同時其最小化判別函數的結果是可以媲美傳統的方法。 / Lai, Sai Yu. / Thesis M.Phil. Chinese University of Hong Kong 2015. / Includes bibliographical references (leaves 31-34). / Abstracts also in Chinese. / Title from PDF title page (viewed on 18, October, 2016). / Detailed summary in vernacular field only.

Cheung, Ka Wai Ricker. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2007. / Includes bibliographical references (leaves 68-69). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Basic Concepts --- p.1 / Chapter 1.2 --- Linkage Disequilibrium --- p.5 / Chapter 1.3 --- Transmission Disequilibrium Test --- p.7 / Chapter 1.4 --- Scope of Thesis --- p.8 / Chapter 2 --- Transmission Disequilibrium Test --- p.9 / Chapter 2.1 --- The Model --- p.9 / Chapter 2.2 --- The Data Structure and The Statistic --- p.12 / Chapter 3 --- Small Sample Properties of Transmission Disequilibrium Test --- p.16 / Chapter 3.1 --- Exact Distribution of TDT Statistic --- p.16 / Chapter 3.2 --- Power under Alternative Hypothesis --- p.20 / Chapter 3.3 --- P-Value --- p.29 / Chapter 4 --- Exact P-Value and Power --- p.35 / Chapter 5 --- Haplotype Relative Risk --- p.61 / Chapter 6 --- Conclusion --- p.66 / References --- p.68

Digitized by Kansas Correctional Industries

This dissertation concerns two topics in the analysis of finite population surveys: setting sample size and hypothesis testing. The first concerns the a priori determination of the sample size needed to obtain species members. The second concerns testing distributional hypotheses when two equal-size populations are sampled. Setting sample size to obtain species is a problem which arises when an investigator wants to obtain (1) a member of all species present in an area (2) a member of all species whose relative frequency is greater than, say, 20% or (3) a member of each species in a target set of species. Chapter 2 presents a practical solution to these questions by setting a target sample size for which the species are obtained with known probability. The solution requires the estimated relative frequency of the rarest species of interest; total number of species is not needed. Because this problem has substantial computational demands, easy-to-compute formulas are needed and given. Three practical examples are presented. Testing of finite population distributional hypotheses is covered in Chapter 3. The test proposed here works under reasonably general designs and is based on a Horvitz-Thompson type correction of the usual Mann-Whitney U statistic. The investigation here compared this proposed test to a corrected (for finiteness) form of the usual Wilcoxon rank sum test. Size and power of the two test procedures are investigated using simulation. The proposed test had approximately correct nominal size over a wide range of situations. The corrected Wilcoxon test exhibited extreme violations in size in many cases. Power of the two tests in situations where they have equal size is similar in most practically interesting cases. / Graduation date: 1996

Habitat loss is one of the leading causes of species decline across all taxa and conservation practices require information on population trends. The Mescalero Sands ecosystem, New Mexico, USA, is experiencing landscape changes associated with oil and gas development. The dune-dwelling lizard community contains a habitat specialist, Sceloporus arenicolus, that is of particular interest because it has a very limited geographic distribution that is entirely subject to oil and gas development. Distance sampling is widely used to estimate population densities of many vertebrate taxa however assumptions can be difficult to satisfy with certain species or in certain habitats. Researchers must investigate the likelihood that assumptions can be satisfied before implementing any population sampling method. I had two objectives. First to investigate the precision of population densities of dune-dwelling lizards estimated via distance sampling that was coupled with double-observer surveys. Second to compare abundances of dune-dwelling lizards among sites that varied in oil and gas development. I conducted distance line transects and compared those density estimates to densities obtained from total removal plots. I quantified the amount of oil and gas development, habitat quantity and quality and correlated those to lizard abundances to investigate the effects of oil and gas development on lizard populations. I found large differences in density estimates from distance sampling and total removal plots that resulted from violation of distance sampling assumptions. Although distance sampling is a low cost method, it does not produce reliable density estimates for dune-dwelling lizards and is not an appropriate sampling method in this system. I did not find oil and gas development effects on the habitat quantity, quality or on the abundances of lizards. Lizard abundances were most strongly correlated to habitat quantity. Lizard abundances may be influenced by complex interactions between oil and gas development and habitat quantity and quality but controlling for those interactions was beyond the scope of my study. Before and after experiments and long-term studies at multiple sites would be required to more fully address the effects of oil and gas development on lizard populations in the Mescalero Sands.

Foreground extraction technology plays an important role in image and video processing tasks. It has been widely used in various industries. To better describe the overlap relationship between foreground and background, alpha channel is introduced. It reveals the opacity property of foreground objects. Thus, fully extracting a foreground object requires determining the alpha values for pixels, also known as extracting an alpha matte. In this thesis, we propose an improved sampling-based alpha matting algorithm, which is capable of generating high quality matting results. By analyzing the weakness of previous approaches, we optimize the sampling process and consider the cost of each sample pair to avoid missing any good samples. The good performance is demonstrated even for complex images. On the other hand, extracting foreground objects from video sequences is a more challenging task since it has higher demands on accuracy and efficiency. Previous approaches usually require a significant amount of user input and the results still suffer from inaccuracy. In this thesis, we successfully extend our algorithm to video sequences and let it run in an automatic fashion. Adaptive trimap, which is vital for matting, can be automatically generated and properly propagated in this system. Our method not only reduces the user interference but also guarantees the matting quality.