Master of Science / Department of Statistics / Weixin Yao / In practice, when applying a statistical method it often occurs that some observations deviate from the usual model assumptions. Least-squares (LS) estimators are very sensitive to outliers. Even one single atypical value may have a large effect on the regression parameter estimates. The goal of robust regression is to develop methods that are resistant to the possibility that one or several unknown outliers may occur anywhere in the data. In this paper, we review various robust regression methods including: M-estimate, LMS estimate, LTS estimate, S-estimate, [tau]-estimate, MM-estimate, GM-estimate, and REWLS estimate. Finally, we compare these robust estimates based on their robustness and efficiency through a simulation study. A real data set application is also provided to compare the robust estimates with traditional least squares estimator.

High quality defense personnel have decisive influence at modern war, and therefore it is the benefit for national security, and the root, garuantee for enhancing military combat power. For years, highly personnel resignation rate has been an important issue for militart personnel resources management. Abnormal resignation rate will not only influences the quality of organizational operation but also disr pts the experience of personnel of the organizational structure.Especially for military services,it will effect our national security and combat power as a whole. General studies of probing resignation were most focuset on factors of resignation will,tendency as probing issues,seldom studies were focused on systematic stuies of resignation rate. Therefore, it is a respond of human resources policies to probe resignation rate in an appropriate way. In this stay, the commissioned and non-commissioned offices in Chinese Marine Corp who stationed in southern Taiwan were taken as probing factors. The predictable capability of Logistic Regression Model has been used in this study as well in order to create the calculation model mode for resignanation rate. The result of the study has been comfirmed that educational level, part-time studies, seniority, marriage, ranking, branch of military services, salary, unit character, welfare and so on were all resignationrelared. Also it is acceptable to predict resignation rate by utilizing this method.

This report presents a top-down methodology to forecast annual international flight operations at sixty-six U.S. airports, whose combined operations accounted for 99.8% of the total international passenger flight operations in National Airspace System (NAS) in 2004. The forecast of international flight operations at each airport is derived from the combination of passenger flight operations at the airport to ten World Regions. The regions include: Europe, Asia, Africa, South America, Mexico, Canada, Caribbean and Central America, Middle East, Oceania and U.S. International. In the forecast, a "top-down" methodology is applied in three steps. In the fist step, individual linear regression models are developed to forecast the total annual international passenger enplanements from the U.S. to each of nine World Regions. The resulting regression models are statistically valid and have parameters that are credible in terms of signs and magnitude. In the second step, the forecasted passenger enplanements are distributed among international airports in the U.S. using individual airport market share factors. The airport market share analysis conducted in this step concludes that the airline business is the critical factor explaining the changes associated with airport market share. In the third and final step, the international passenger enplanements at each airport are converted to flight operations required for transporting the passengers. In this process, average load factor and average seats per aircraft are used. The model has been integrated into the Transportation Systems Analysis Model (TSAM), a comprehensive intercity transportation planning tool. Through a simple graphic user interface implemented in the TSAM model, the user can test different future scenarios by defining a series of scaling factors for GDP, load factor and average seats per aircraft. The default values for the latter two variables are predefined in the model using 2004 historical data derived from Department of Transportation T100 international segment data. / Master of Science

Optimal experimental designs for generalized linear models have received increasing attention in recent years. Yet, most of the current research focuses on binary data models especially the one-variable first-order logistic regression model. This research extends this topic to count data models. The primary goal of this research is to develop efficient and robust experimental designs for the Poisson regression model in toxicity studies. D-optimal designs for both the one-toxicant second-order model and the two-toxicant interaction model are developed and their dependence upon the model parameters is investigated. Application of the D-optimal designs is very limited due to the fact that these optimal designs, in terms of ED levels, depend upon the unknown parameters. Thus, some practical designs like equally spaced designs and conditional D-optimal designs, which, in terms of ED levels, are independent of the parameters, are studied. It turns out that these practical designs are quite efficient when the design space is restricted. Designs found in terms of ED levels like D-optimal designs are not robust to parameters misspecification. To deal with this problem, sequential designs are proposed for Poisson regression models. Both fully sequential designs and two-stage designs are studied and they are found to be efficient and robust to parameter misspecification. For experiments that involve two or more toxicants, restrictions on the survival proportion lead to restricted design regions dependent on the unknown parameters. It is found that sequential designs perform very well under such restrictions. In most of this research, the log link is assumed to be the true link function for the model. However, in some applications, more than one link functions fit the data very well. To help identify the link function that generates the data, experimental designs for discrimination between two competing link functions are investigated. T-optimal designs for discrimination between the log link and other link functions such as the square root link and the identity link are developed. To relax the dependence of T-optimal designs on the model truth, sequential designs are studied, which are found to converge to T-optimal designs for large experiments. / Ph. D.

The purpose of this study was to develop a regression equation to predict VO2max based on non-exercise (N-EX) data. All participants (N = 100), aged 18-65 years old, successfully completed a maximal graded exercise test (GXT) to assess VO2max (mean ± SD; 39.96 mL∙kg-¹∙min&sup-1; ± 9.54 mL∙kg-¹∙min-¹). The N-EX data collected just before the maximal GXT included the participant's age, gender, body mass index (BMI), perceived functional ability (PFA) to walk, jog, or run given distances, and current physical activity (PA-R) level. Multiple linear regression generated the following N-EX prediction equation (R = .93, SEE = 3.45 mL∙kg-¹∙min-¹, %SEE = 8.62): VO2max (mL∙kg-¹∙min-¹) = 48.0730 + (6.1779 x gender) - (0.2463 x age) - (0.6186 x BMI) + (0.7115 x PFA) + (0.6709 x PA-R). Cross validation using PRESS (predicted residual sum of squares) statistics revealed minimal shrinkage (Rp = .91 and SEEp = 3.63 mL∙kg-¹∙min-¹); thus, this model should yield acceptable accuracy when applied to an independent sample of adults (aged 18-65) with a similar cardiorespiratory fitness level. Based on standardized β-weights the PFA variable (0.41) was the most effective at predicting VO2max followed by age (-0.34), gender (0.33), BMI (-0.27), and PA-R (0.16). This study provides a N-EX regression model that yields relatively accurate results and is a convenient way to predict VO2max in adult men and women.

The level of labor productivity differs among the European Union countries, especially when you compare a developing country to a more developed country in the EU. This is an issue because the achievement of high labor productivity is a necessary stipulation for a developing economy to realize economic growth and more economic development. On the other hand, the more individuals in an economy with access to the internet (internet connectivity) depicts how developed the economy is in terms of information and communication technology (ICT). Accordingly, the purpose of this paper is to ascertain whether there is a positive relationship between countries having high internet connectivity and labor productivity in the EU. In doing so, Political and entrepreneurial decision-makers can use these findings to decide how much attention or budget to put on the ICT sector to improve labor productivity. To understand the factors that affect labor productivity, Adam Smith and Karl Marx’s theory on labor productivity is used to gain a better understanding. A panel data analysis using a fixed-effect model and pooled OLS regression model is applied in the study to predict the relationship. The result of the study indicates that internet connectivity does not have a significant impact on Labour productivity, or there was not enough evidence showing that they are positively correlated with each other.

Master of Science / Department of Statistics / Weixing Song / A nonparametric lack-of-fit test is proposed to check the adequacy of the presumed parametric form for the regression function in Tobit regression models by applying Zheng's device with weighted residuals. It is shown that testing the null hypothesis for the standard Tobit regression models is equivalent to test a new null hypothesis of the classic regression models. An optimal weight function is identified to maximize the local power of the test. The test statistic proposed is shown to be asymptotically normal under null hypothesis, consistent against some fixed alternatives, and has nontrivial power for some local nonparametric power for some local nonparametric alternatives. The finite sample performance of the proposed test is assessed by Monte-Carlo simulations. An empirical study is conducted based on the data of University of Michigan Panel Study of Income Dynamics for the year 1975.

There has been little research conducted on the physiological response of calves to temperatures outside thermal neutrality and its effects on intake and weight gain. The effects of ambient temperature on Holstein dairy calves intakes and weight gain were evaluated over a 12-month period. Ambient temperature was monitored using a weather station located 1.3 kilometers from the Utah State University Caine Dairy. Calf health was monitored daily using the University of Wisconsin-Madison School of Veterinary Medicine scoring criteria. Calves were fed whole milk and free choice calf starter. Weight gain, hip height, starter intake, and weather data (temperature, wind speed, relative humidity, precipitation, and barometric pressure) were averaged for 7-day intervals beginning at birth through 13 weeks of age. A regression model was developed including starter intake, milk intake, hip and wither height, calf heath scores, and weather data with weight gain as the dependent variable for each of the 4 seasons of the year. The fall season (September, October, and November) had a negative impact on calf intake and weight gain (averaging 20 pounds (9.1 kilograms) less at 2 months) than other seasons. Calves raised in the winter months also ate significantly more starter, but had the same weight gain as other seasons. Environmental stress factors impact animal welfare and animal productivity, which in turn impacts the economics of the dairy operation and should also be used in determining husbandry practices.

Some highway agencies in the United States are experiencing frequent distresses in asphalt pavements on bridge approaches/departures. Commonly observed distresses include alligator cracking and rutting, which reduce roadway smoothness and safety. To lessen the distresses in pavements it is needed to investigate the extent and root causes of the problem. Based on Florida highway conditions, this research study mainly focused on1. Literature review and identification of the extent of the problem; 2. Collection of relevant pavement condition data and descriptive analysis; 3. Development of statistical models to determine factors influencing the distresses in asphalt pavements on bridge approaches/departures. To the best of my knowledge, this is the first study that uses a statistical model to determine the factors that are responsible for causing asphalt pavement distresses on bridge approaches/departures. As part of the literature review, a nationwide questionnaire survey was targeted towards U.S state DOTs. The data collection and analysis specific to the Florida highways found that in 2015 on Florida Interstate highways, about 27% bridges with asphalt pavements on their approaches/departures showed signs of cracking, and about 20% bridges have noticeable rutting in their approach or departure pavements. A random parameter linear regression model was applied to examine the factors that may influence distresses in asphalt pavements in Florida. Pavement condition was evaluated based on the Florida Department of Transportation (FDOT) 2015 pavement condition data and video log images, and other relevant data were collected from various sources such as FDOT Roadway Characteristics Inventory (RCI) database, FDOT pavement management reports, and FDOT Ground Penetrating Radar (GPR) survey reports. A constraint existed in the availability of the GPR data that can give pavement layer thickness, which limited the number of bridge approach pavement sections included in the statistical modeling. Based on the limited data, the estimated results from the random parameter linear regression model showed that the variables influencing distresses in asphalt pavements on bridge approaches/departures, in terms of rutting and roughness, may include pavement age, annual average daily truck traffic, and surface friction course.

In this project, two broad facets in the design of a methodology for performance optimization of indexable carbide inserts were examined. They were physical destructive testing and software simulation.For the physical testing, statistical research techniques were used for the design of the methodology. A five step method which began with Problem definition, through System identification, Statistical model formation, Data collection and Statistical analyses and results was indepthly elaborated upon. Set-up and execution of an experiment with a compression machine together with roadblocks and possible solution to curb road blocks to quality data collection were examined. 2k factorial design was illustrated and recommended for process improvement. Instances of first-order and second-order response surface analyses were encountered. In the case of curvature, test for curvature significance with center point analysis was recommended. Process optimization with method of steepest ascent and central composite design or process robustness studies of response surface analyses were also recommended.For the simulation test, AdvantEdge program was identified as the most used software for tool development. Challenges to the efficient application of this software were identified and possible solutions proposed. In conclusion, software simulation and physical testing were recommended to meet the objective of the project.