Applied Statistics Commons^™

Generalizing Multistage Partition Procedures For Two-Parameter Exponential Populations, Rui Wang

LSU New Orleans Theses and Dissertations

ANOVA analysis is a classic tool for multiple comparisons and has been widely used in numerous disciplines due to its simplicity and convenience. The ANOVA procedure is designed to test if a number of different populations are all different. This is followed by usual multiple comparison tests to rank the populations. However, the probability of selecting the best population via ANOVA procedure does not guarantee the probability to be larger than some desired prespecified level. This lack of desirability of the ANOVA procedure was overcome by researchers in early 1950's by designing experiments with the goal of selecting the best …

Development Of A Statistical Model For Discrimination Of Rupture Status In Posterior Communicating Artery Aneurysms, Felicitas J. Detmer, Bong Jae Chung, Fernando Mut, Michael Pritz, Martin Slawski, Farid Hamzei-Sichani, David Kallmes, Christopher Putman, Carlos Jimenez, Juan R. Cebral

Department of Applied Mathematics and Statistics Faculty Scholarship and Creative Works

Background: Intracranial aneurysms at the posterior communicating artery (PCOM) are known to have high rupture rates compared to other locations. We developed and internally validated a statistical model discriminating between ruptured and unruptured PCOM aneurysms based on hemodynamic and geometric parameters, angio-architectures, and patient age with the objective of its future use for aneurysm risk assessment. Methods: A total of 289 PCOM aneurysms in 272 patients modeled with image-based computational fluid dynamics (CFD) were used to construct statistical models using logistic group lasso regression. These models were evaluated with respect to discrimination power and goodness of fit using tenfold nested …

Bayesian Analytical Approaches For Metabolomics : A Novel Method For Molecular Structure-Informed Metabolite Interaction Modeling, A Novel Diagnostic Model For Differentiating Myocardial Infarction Type, And Approaches For Compound Identification Given Mass Spectrometry Data., Patrick J. Trainor

Electronic Theses and Dissertations

Metabolomics, the study of small molecules in biological systems, has enjoyed great success in enabling researchers to examine disease-associated metabolic dysregulation and has been utilized for the discovery biomarkers of disease and phenotypic states. In spite of recent technological advances in the analytical platforms utilized in metabolomics and the proliferation of tools for the analysis of metabolomics data, significant challenges in metabolomics data analyses remain. In this dissertation, we present three of these challenges and Bayesian methodological solutions for each. In the first part we develop a new methodology to serve a basis for making higher order inferences in metabolomics, …

Comparison Of Correlation, Partial Correlation, And Conditional Mutual Information For Interaction Effects Screening In Generalized Linear Models, Ji Li

Graduate Theses and Dissertations

Numerous screening techniques have been developed in recent years for genome-wide association studies (GWASs) (Moore et al., 2010). In this thesis, a novel model-free screening method was developed and validated by an extensive simulation study. Many screening methods were mainly focused on main effects, while very few studies considered the models containing both main effects and interaction effects. In this work, the interaction effects were fully considered and three different methods (Pearson’s Correlation Coefficient, Partial Correlation, and Conditional Mutual Information) were tested and their prediction accuracies were compared.

Pearson’s Correlation Coefficient method, which is a direct interaction screening (DIS) procedure, …

Clustering Mixed Data: An Extension Of The Gower Coefficient With Weighted L2 Distance, Augustine Oppong

Electronic Theses and Dissertations

Sorting out data into partitions is increasing becoming complex as the constituents of data is growing outward everyday. Mixed data comprises continuous, categorical, directional functional and other types of variables. Clustering mixed data is based on special dissimilarities of the variables. Some data types may influence the clustering solution. Assigning appropriate weight to the functional data may improve the performance of the clustering algorithm. In this paper we use the extension of the Gower coefficient with judciously chosen weight for the L2 to cluster mixed data.The benefits of weighting are demonstrated both in in applications to the Buoy data set …

Pretrial Release And Failure-To-Appear In Mclean County, Il, Jonathan Monsma

Student Research – Stevenson Center

Actuarial risk assessment tools increasingly have been employed in jurisdictions across the U.S. to assist courts in the decision of whether someone charged with a crime should be detained or released prior to their trial. These tools should be continually monitored and researched by independent 3^rd parties to ensure that these powerful tools are being administered properly and used in the most proficient way as to provide socially optimal results. McLean County, Illinois began using the Public Safety Assessment-Court^TM (PSA-Court or simply PSA) risk assessment tool beginning in 2016. This study culls data from the McLean County Jail …

A Distance Based Method For Solving Multi-Objective Optimization Problems, Murshid Kamal, Syed Aqib Jalil, Syed Mohd Muneeb, Irfan Ali

Journal of Modern Applied Statistical Methods

A new model for the weighted method of goal programming is proposed based on minimizing the distances between ideal objectives to feasible objective space. It provides the best compromised solution for Multi Objective Linear Programming Problems (MOLPP). The proposed model tackles MOLPP by solving a series of single objective sub-problems, where the objectives are transformed into constraints. The compromise solution so obtained may be improved by defining priorities in terms of the weight. A criterion is also proposed for deciding the best compromise solution. Applications of the algorithm are discussed for transportation and assignment problems involving multiple and conflicting objectives. …

Goalie Analytics: Statistical Evaluation Of Context-Specific Goalie Performance Measures In The National Hockey League, Marc Naples, Logan Gage, Amy Nussbaum

SMU Data Science Review

In this paper, we attempt to improve upon the classic formulation of save percentage in the NHL by controlling the context of the shots and use alternative measures than save percentage. In particular, we find save percentage to be both a weakly repeatable skill and predictor of future performance, and we seek other goalie performance calculations that are more robust. To do so, we use three primary tests to test intra-season consistency, intra-season predictability, and inter-season consistency, and extend the analysis to disentangle team effects on goalie statistics. We find that there are multiple ways to improve upon classic save …

Data Scientist’S Analysis Toolbox: Comparison Of Python, R, And Sas Performance, Jim Brittain, Mariana Cendon, Jennifer Nizzi, John Pleis

SMU Data Science Review

A quantitative analysis will be performed on experiments utilizing three different tools used for Data Science. The analysis will include replication of analysis along with comparisons of code length, output, and results. Qualitative data will supplement the quantitative findings. The conclusion will provide data support guidance on the correct tool to use for common situations in the field of Data Science.

Estimation Of Finite Population Mean By Using Minimum And Maximum Values In Stratified Random Sampling, Umer Daraz, Javid Shabbir, Hina Khan

Journal of Modern Applied Statistical Methods

In this paper we have suggested an improved class of ratio type estimators in estimating the finite population mean when information on minimum and maximum values of the auxiliary variable is known. The properties of the suggested class of estimators in terms of bias and mean square error are obtained up to first order of approximation. Two data sets are used for efficiency comparisons.

A Bayesian Beta-Mixture Model For Nonparametric Irt (Bbm-Irt), Ethan A. Arenson, George Karabatsos

Journal of Modern Applied Statistical Methods

Item response models typically assume that the item characteristic (step) curves follow a logistic or normal cumulative distribution function, which are strictly monotone functions of person test ability. Such assumptions can be overly-restrictive for real item response data. A simple and more flexible Bayesian nonparametric IRT model for dichotomous items is introduced, which constructs monotone item characteristic (step) curves by a finite mixture of beta distributions, which can support the entire space of monotone curves to any desired degree of accuracy. An adaptive random-walk Metropolis-Hastings algorithm is proposed to estimate the posterior distribution of the model parameters. The Bayesian IRT …

Robust Estimation And Inference On Current Status Data With Applications To Phase Iv Cancer Trial, Deo Kumar Srivastava, Liang Zhu, Melissa M. Hudson, Jianmin Pan, Shesh N. Rai

Journal of Modern Applied Statistical Methods

The use of piecewise exponential distributions was proposed by Rai et al. (2013) for analyzing cardiotoxicity data. Some parametric models are proposed, but the focus is on the Weibull distribution, which overcomes the limitation of piecewise exponential.

Robust Heteroscedasticity Consistent Covariance Matrix Estimator Based On Robust Mahalanobis Distance And Diagnostic Robust Generalized Potential Weighting Methods In Linear Regression, M. Habshah, Muhammad Sani, Jayanthi Arasan

Journal of Modern Applied Statistical Methods

The violation of the assumption of homoscedasticity and the presence of high leverage points (HLPs) are common in the use of regression models. The weighted least squares can provide the solution to heteroscedastic regression model if the heteroscedastic error structures are known. Based on Furno (1996), two robust weighting methods are proposed based on HLP detection measures (robust Mahalanobis distance based on minimum volume ellipsoid and diagnostic robust generalized potential based on index set equality (DRGP(ISE)) on robust heteroscedasticity consistent covariance matrix estimators. Results obtained from a simulation study and real data sets indicated the DRGP(ISE) method is superior.

Masked Instability: Within-Sector Financial Risk In The Presence Of Wealth Inequality, Youngna Choi

Department of Applied Mathematics and Statistics Faculty Scholarship and Creative Works

We investigate masked financial instability caused by wealth inequality. When an economic sector is decomposed into two subsectors that possess a severe wealth inequality, the sector in entirety can look financially stable while the two subsectors possess extreme financially instabilities of opposite nature, one from excessive equity, the other from lack thereof. The unstable subsector can result in further financial distress and even trigger a financial crisis. The market instability indicator, an early warning system derived from dynamical systems applied to agent-based models, is used to analyze the subsectoral financial instabilities. Detailed mathematical analysis is provided to explain what financial …

Internal Consistency Reliability In Measurement: Aggregate And Multilevel Approaches, Georgios Sideridis, Abdullah Saddaawi, Khaleel Al-Harbi

Journal of Modern Applied Statistical Methods

The purpose of the present paper was to evaluate the internal consistency reliability of the General Teacher Test assuming clustered and non-clustered data using commercial software (Mplus). Participants were 2,000 testees who were selected using random sampling from a larger pool of examinees (more than 65k). The measure involved four factors, namely: (a) planning for learning, (b) promoting learning, (c) supporting learning, and (d) professional responsibilities and was hypothesized to comprise a unidimensional instrument assessing generalized skills and competencies. Intra-class correlation coefficients and variance ratio statistics suggested the need to incorporate a clustering variable (i.e., university) when evaluating the factor …

Fitting The Rasch Model Under The Logistic Regression Framework To Reduce Estimation Bias, Tianshu Pan

Journal of Modern Applied Statistical Methods

This article showed how and why the Rasch model can be fitted under the logistic regression framework. Then a penalized maximum likelihood (Firth 1993) for logistic regression models can also be used to reduce ML biases when fitting the Rasch model. These conclusions are supported by a simulation study.

Regressions Regularized By Correlations, Stan Lipovetsky

Journal of Modern Applied Statistical Methods

The regularization of multiple regression by proportionality to correlations of predictors with dependent variable is applied to the least squares objective and normal equations to relax the exact equalities and to get a robust solution. This technique produces models not prone to multicollinearity and is very useful in practical applications.

An Explanatory Study On The Non-Parametric Multivariate T2 Control Chart, Abdolrasoul Mostajeran, Nasrolah Iranpanah, Rassoul Noorossana

Journal of Modern Applied Statistical Methods

Most control charts require the assumption of normal distribution for observations. When distribution is not normal, one can use non-parametric control charts such as sign control chart. A deficiency of such control charts could be the loss of information due to replacing an observation with its sign or rank. Furthermore, because the chart statistics of T² are correlated, the T² chart is not a desire performance. Non-parametric bootstrap algorithm could help to calculate control chart parameters using the original observations while no assumption regarding the distribution is needed. In this paper, first, a bootstrap multivariate control chart is …

Optimum Stratification In Bivariate Auxiliary Variables Under Neyman Allocation, Faizan Danish, S.E.H. Rizvi

Journal of Modern Applied Statistical Methods

In several situations complete data set of the study variable is unknown that becomes a stumbling block in various stratification techniques in order to obtain stratification points on two way stratification method. In this paper a technique has been proposed under Neyman allocation when the stratification is done oj the two auxiliary variable having one estimation variable under consideration. Due to complexities created by minimal equations approximate optimum strata boundaries has been obtained. Empirical study has been done to illustrate the proposed method when the auxiliary variables have standard Cauchy and power distributions.

Estimation Of Zero-Inflated Population Mean: A Bootstrapping Approach, Khyam Paneru, R. Noah Padgett, Hanfeng Chen

Journal of Modern Applied Statistical Methods

A mixture model was adopted from the maximum pseudo-likelihood approach under complex sampling designs to estimate the mean of zero-inflated population. To overcome the complexity and assumptions of asymptotic distribution, the maximum pseudo-likelihood function was used, but a bootstrapping procedure was proposed as an alternative. Bootstrap confidence intervals consistently capture the true means of zero-inflated populations of the simulation studies.

Moment Generating Functions Of Complementary Exponential-Geometric Distribution Based On K-Th Lower Record Values, Devendra Kumar, Sanku Dey, Mansoor Rashid Malik, Fahad M. Al-Aboud

Journal of Modern Applied Statistical Methods

The complementary exponential-geometric (CEG) distribution is a useful model for analyzing lifetime data. For this distribution, some recurrence relations satisfied by marginal and joint moment generating functions of k-th lower record values were established. They enable the computation of the means, variances, and covariances of k-th lower record values for all sample sizes in a simple and efficient recursive manner. Means, variances, and covariances of lower record values were tabulated from samples of sizes up to 10 for various values of the parameters.

Optimal Model Selection For Truncated Data Among Non-Nested Competitive Models, Parisa Torkaman

Journal of Modern Applied Statistical Methods

Selecting a model for incomplete data is an important issue. Truncated data is an example of incomplete data, which sometimes occurs due to inherent limitations. The maximum likelihood estimator features and its asymptotic distribution are studied, and a test statistic among non-nested competitive model of incomplete data is presented, which can select an appropriate model close to the true model. This close-to-true model under the null hypothesis of the equivalency of two competitive models against alternative hypothesis is selected.

Letter To The Editor: Regarding A Possible Non-Null Interpretation Of The Michelson-Morley Experiment, Maurizio Consoli

Journal of Modern Applied Statistical Methods

The author writes in response to Sawilowsky in JMASM 2(2) and 4(1).

The Transmuted Exponentiated Additive Weibull Distribution: Properties And Applications, Zohdy M. Nofal, Ahmed Z. Afify, Haitham M. Yousof, Daniele Cristina Tita Granzotto, Francisco Louzada

Journal of Modern Applied Statistical Methods

A new generalization of the transmuted additive Weibull distribution is proposed by using the quadratic rank transmutation map, the so-called transmuted exponentiated additive Weibull distribution. It retains the characteristics of a good model. It is more flexible, being able to analyze more complex data; it includes twenty-seven sub-models as special cases and it is interpretable. Several mathematical properties of the new distribution as closed forms for ordinary and incomplete moments, quantiles, and moment generating function are presented, as well as the MLEs. The usefulness of the model is illustrated by using two real data sets.

Modeling Insurance Claims Using Flexible Skewed And Mixture Probability Distributions, Aaron J. Leinwander, Mohammad A. Aziz

Journal of Modern Applied Statistical Methods

The normal distribution comes as a first choice when fitting real data, but it may not be suitable if the assumed distribution deviates from normality. Flexible skewed distributions are capable of including skewness and taking into account multimodality. They may be applied to find appropriate distributions for describing the claim amounts in insurance. The objective is to model insurance claims using a set of flexible skewed and mixture probability distributions, and to test how well they fit the claims. Results indicate the skew-t distribution and alpha-skew Laplace distribution are able to describe unimodal claims accurately, whereas scale mixture of …

Single Missing Data Imputation In Pls-Based Structural Equation Modeling, Ned Kock

Journal of Modern Applied Statistical Methods

Missing data, a source of bias in structural equation modeling (SEM) employing the partial least squares method (PLS), are commonly handled with deletion methods such as listwise and pairwise deletion. Missing data imputation methods do not resort to deletion. Five single missing data imputation methods are considered employing the PLS Mode A algorithm of which two hierarchical methods are new. The results of a Monte Carlo experiment suggest that Multiple Regression Imputation yielded the least biased mean path coefficient estimates, followed by Arithmetic Mean Imputation. With respect to mean loading estimates, Arithmetic Mean Imputation yielded the least biased results, followed …

A New Lifetime Distribution For Series System: Model, Properties And Application, Adil Rashid, Zahooor Ahmad, T R. Jan

Journal of Modern Applied Statistical Methods

A new lifetime distribution for modeling system lifetime in series setting is proposed that embodies most of the compound lifetime distribution. The reliability analysis of parent and of sub-models has also been discussed. Various mathematical properties that include moment generating function, moments, and order statistics have been obtained. The newly-proposed distribution has a flexible density function; more importantly its hazard rate function can take up different shapes such as bathtub, upside down bathtub, increasing, and decreasing shapes. The unknown parameters of the proposed generalized family have been estimated through MLE technique. The strength and usefulness of the proposed family was …

An Inferential Method For Determining Which Of Two Independent Variables Is Most Important When There Is Curvature, Rand Wilcox

Journal of Modern Applied Statistical Methods

Consider three random variables Y, X₁ and X₂, where the typical value of Y, given X₁ and X₂, is given by some unknown function m(X₁, X₂). A goal is to determine which of the two independent variables is most important when both variables are included in the model. Let τ₁ denote the strength of the association associated with Y and X₁, when X₂ is included in the model, and let τ₂ be defined in an analogous manner. If it is assumed …

Sample Size For Non-Inferiority Tests For One Proportion: A Simulation Study, Özlem Güllü, Mustafa Agah Tekindal

Journal of Modern Applied Statistical Methods

The objective of non-inferiority trials is to demonstrate the efficiency of a novel treatment whether it is acceptably less or more efficient than a control or active (existing) treatment. They are employed in situations where, when compared to the active treatment, the novel treatment is to be advantageous with higher rates of reliability, compatibility, cost-efficiency, etc. Odds ratio is the most significant measure used in investigating the size of efficiency of treatments relative to one another. The purpose of the study is to calculate and evaluate the sample size under different scenarios based on three different test statistics in non-inferiority …

Handling Missing Data In Single-Case Studies, Chao-Ying Joanne Peng, Li-Ting Chen

Journal of Modern Applied Statistical Methods

Multiple imputation is illustrated for dealing with missing data in a published SCED study. Results were compared to those obtained from available data. Merits and issues of implementation are discussed. Recommendations are offered on primal/advanced readings, statistical software, and future research.