Search

Profiling analysis aims to evaluate health care providers, including hospitals, nursing homes, or dialysis facilities among others with respect to a patient outcome, such as 30-day unplanned hospital readmission or mortality. Fixed effects (FE) profiling models have been developed over the last decade, motivated by the overall need to (a) improve accurate identification or flagging of under-performing providers, (b) relax assumptions inherent in random effects (RE) profiling models, and (c) take into consideration the unique disease characteristics and care/treatment processes of end-stage kidney disease (ESKD) patients on dialysis. In this paper, we review the current state of FE methodologies and their rationale in the ESKD population and illustrate applications in four key areas: profiling dialysis facilities for (1) patient hospitalizations over time (longitudinally) using standardized dynamic readmission ratio (SDRR), (2) identification of dialysis facility characteristics (e.g., staffing level) that contribute to hospital readmission, and (3) adverse recurrent events using standardized event ratio (SER). Also, we examine the operating characteristics with a focus on FE profiling models. Throughout these areas of applications to the ESKD population, we identify challenges for future research in both methodology and clinical studies.

The generalized linear mixed model for binary outcomes with the probit link function is used in many fields but has a computationally challenging likelihood when there are many random effects. We extend a previously used importance sampler, making it much faster in the context of estimating heritability and related effects from family data by adding a gradient and a Hessian approximation and making a faster implementation. Additionally, a graph-based method is suggested to simplify the likelihood when there are thousands of individuals in each family. Simulation studies show that the resulting method is orders of magnitude faster, has a negligible efficiency loss, and confidence intervals with nominal coverage. We also analyze data from a large study of obsessive-compulsive disorder based on Swedish multi-generational data. In this analysis, the proposed method yielded similar results to a previous analysis, but was much faster. Supplementary materials for this article are available online., QC 20240807

Background and Aims: In systematic studies the use of anatomical characters is of great relevance to support or refute the circumscription of genera and species. Nahuatlea is a genus of the tribe Gochnatieae (Asteraceae), and includes six species, supported by molecular evidence. However, its anatomical characters have not yet been evaluated from a taxonomic point of view. The objective of this work was to describe the leaf architecture and anatomy as well as the wood anatomy of Nahuatlea species and its sister groups Tehuasca, Anastraphia and Cnicothamnus to evaluate and record attributes with possible systematic value.Methods: Wood and leaf samples from three individuals per species were analyzed using conventional microtechnical techniques to obtain and de-scribe qualitative and quantitative characters. Characters were analyzed using generalized linear mixed models and cluster analysis.Key results: Characteristics of the wood such as porosity, arrangement of the vessels, aperture of intervessel pits and number of series in the rays present possible taxonomic value. For example, vessels with a diagonal pattern are unique to N. hypoleuca, while bi- or triseriate rays are unique to species of Nahuatlea. In the leaf, the venation patterns and the occurrence of hypodermis sheath extensions present possible taxonomic value. The branching veinlets and hypodermis of one layer are exclusive to N. purpusii. The analyses showed significant differences for some of the characters of the wood and the leaf.Conclusions: The anatomy of wood and leaf allows the recognition of unique combinations of characters that support the delimitation of Nahuatleaspecies., Antecedentes y Objetivos: En los estudios sistemáticos el uso de los caracteres anatómicos es de gran relevancia para apoyar o refutar la circunscrip-ción de géneros y especies. Nahuatlea es un género de la tribu Gochnatieae (Asteraceae), que incluye seis especies sustentadas con evidencia mole-cular. Sin embargo, sus caracteres anatómicos aún no han sido evaluados desde el punto de vista taxonómico. El objetivo de este trabajo fue describir la arquitectura y anatomía foliar, así como la anatomía de la madera de las especies de Nahuatlea y sus grupos hermanos Tehuasca, Anastraphia y Cnicothamnus para evaluar e identificar atributos con posible valor sistemático.Métodos:Se analizaronmuestras de madera y hojas de tres individuos por especie, mediante técnicas convencionales de microtecniapara obtener y describir los caracteres cualitativos y cuantitativos. Estos se analizaron mediante modelos mixtos lineares generalizados y análisis de agrupación aglomerativa. Resultados clave: Caracteres de la madera como porosidad, disposición de los vasos, apertura de las punteaduras intervasculares y número de series en los radios presentan posible valor taxonómico. Por ejemplo, vasos con un patrón diagonal son exclusivos de N. hypoleuca, mientras que los radios bi o triseriados lo son para todas las especies de Nahuatlea. En la hoja, los patrones de venación y la presencia de extensiones de vaina e hipodermis presentan posible valor taxonómico. Las vénulas ramificadas e hipodermis de un estrato son exclusivos de N. purpusii. Los análisis mostraron diferen-cias significativas para algunos caracteres de la madera y la hoja. Conclusiones: La anatomía de madera y hoja permite reconocer combinaciones únicas de caracteres que apoyan la delimitación de las especies de Nahuatlea

Profiling analysis aims to evaluate health care providers, including hospitals, nursing homes, or dialysis facilities among others with respect to a patient outcome, such as 30-day unplanned hospital readmission or mortality. Fixed effects (FE) profiling models have been developed over the last decade, motivated by the overall need to (a) improve accurate identification or “flagging” of under-performing providers, (b) relax assumptions inherent in random effects (RE) profiling models, and (c) take into consideration the unique disease characteristics and care/treatment processes of end-stage kidney disease (ESKD) patients on dialysis. In this paper, we review the current state of FE methodologies and their rationale in the ESKD population and illustrate applications in four key areas: profiling dialysis facilities for (1) patient hospitalizations over time (longitudinally) using standardized dynamic readmission ratio (SDRR), (2) identification of dialysis facility characteristics (e.g., staffing level) that contribute to hospital readmission, and (3) adverse recurrent events using standardized event ratio (SER). Also, we examine the operating characteristics with a focus on FE profiling models. Throughout these areas of applications to the ESKD population, we identify challenges for future research in both methodology and clinical studies.

Acord transformatiu CRUE-CSIC, Open Access Funding provided by Universitat Autònoma de Barcelona. This work was partially funded by the grants RTI2018-096072-B-I00 and PID2019-104830RB-I00/AEI DOI:10.13039/501100011033 from the Spanish Ministry of Science, Innovation and Universities, and the Severo Ochoa and María de Maeztu Program for Centers and Units of Excellence in R &D (CEX2020-001084-M) from the Spanish Research Agency., The aim of the current study is to analyze the effects of red and yellow cards on the scoring rate in elite soccer. The sample was composed of 1826 matches in the top five European leagues. All events were structured in 5-min intervals and were analyzed by means of a Generalized Linear Mixed Model with Poisson distribution, considering the presence of correlated data, where the dependent variable is represented by scoring rate. Team strength and home advantage were considered implicitly by means of a transformation of the betting odds for each game. The model also took into account the goal difference and time evolution. Overall, we found that after a sending off, each team's scoring rate changes significantly, damaging the penalised team and favouring its opponent. When the player who is sent off belongs to the Away team, the impact of a red card is more or less maintained over time intervals. The red card effect, on the other hand, tends to fade over time when the affected team is stronger. The relative difference in scoring rates is also affected by the goal difference and the difference in booked players, being slightly lower for the team going ahead if it has more booked players. Our approach allows estimating the expected cumulative soring rate through time for various red card scenarios. Particularly if a red card is given with 30 min of remaining time, the expected impact is 0.39 goals if the guilty player is on the visiting team and 0.50 if he plays for the home team. Coaches and analysts could use this information to establish objectives for players and teams in training and matches and to be prepared for these very different scenarios of numerical superiority or inferiority.

SARS-CoV-2 epidemics quickly propagated worldwide, sorting virus genomic variants in newly established propagules of infections. Stochasticity in transmission within and between countries or an actual selective advantage could explain the global high frequency reached by some genomic variants. Using statistical analyses, demographic reconstructions, and molecular dynamics simulations , we show that the globally invasive G614 spike variant i) underwent a significant demographic expansion in most countries not explained by stochastic effects nor by overrepresentation in clinical samples; ii) increases the spike S1/S2 furin-like site conformational plasticity (short-range effect), and iii) modifies the internal motion of the receptor-binding domain affecting its cross-connection with other functional domains (long-range effect). Our results support the hypothesis of a selective advantage at the basis of the spread of the G614 variant, which we suggest may be due to structural modification of the spike protein at the S1/S2 proteolytic site, and provides structural information to guide the design of variant-specific drugs.

Standardized relative abundance indices for swordfish (Xiphias gladius Linnaeus, 1758) caught by the Spanish surface longline in the western Mediterranean Sea were estimated for the period 1988-2018. Standardized CPUEs in number were estimated through a General Linear Mixed Modeling (GLMM) approach under a negative binomial (NB) error distribution assumption. Standardized CPUEs in biomass were estimated through a General Linear Mixed Modeling (GLMM) approach under a log-normal error distribution assumption. The main factors in the standardization analysis were fishing area and time of the year (quarter). The standardized indices showed notable annual fluctuations without any definite trend for the period under study

Standardized relative abundance indices for swordfish (Xiphias gladius Linnaeus, 1758) caught by the Spanish surface longline in the western Mediterranean Sea were estimated for the period 1988-2018. Standardized CPUEs in number were estimated through a General Linear Mixed Modeling (GLMM) approach under a negative binomial (NB) error distribution assumption. Standardized CPUEs in biomass were estimated through a General Linear Mixed Modeling (GLMM) approach under a log-normal error distribution assumption. The main factors in the standardization analysis were fishing area and time of the year (quarter). The standardized indices showed notable annual fluctuations without any definite trend for the period under study.

Background: Longitudinal randomized controlled trials (RCTs) often aim to test and measure the effect of treatment between arms at a single time point. A two-sample χ2 test is a common statistical approach when outcome data are binary. However, only complete outcomes are used in the analysis. Missing responses are common in longitudinal RCTs and by only analyzing complete data, power may be reduced and estimates could be biased. Generalized linear mixed models (GLMM) with a random intercept can be used to test and estimate the treatment effect, which may increase power and reduce bias. Methods: We simulated longitudinal binary data from a RCT to compare the performance of a complete case χ2 test to a GLMM in terms of power, relative bias, and coverage under different missing data mechanisms. We considered how the baseline probability of the event, within subject correlation, and dropout rates under various missing mechanisms impacted each performance measure. Results: When data were missing completely at random, both χ2 and GLMM produced unbiased estimates; however, the GLMM returned an absolute power gain up to from 12.0% as compared to the χ2 test. When outcome data were missing at random, the GLMM yielded an absolute power gain up to 42.7% and estimates were unbiased or less biased compared to the χ2 test. Conclusions: Investigators wishing to test for a treatment effect between treatment arms in longitudinal RCTs with binary outcome data in the presence of missing data should use a GLMM to gain power and produce minimally unbiased estimates instead of a complete case χ2 test.

Temporal and spatial variation in phenotypic selection due to changing environmental conditions is of great interest to evolutionary biologists, but few existing methods estimating its magnitude take into account the temporal autocorrelation. We use state-space models (SSMs) to analyse phenotypic selection processes that cannot be observed directly and use Template Model builder (TMB), an R package for computing and maximizing the Laplace approximation of the marginal likelihood for SSM and other complex, nonlinear latent variable model via automatic differentiation. Using a long-term great tit (Parus major) dataset, we fit several SSMs and conduct model selection based on Akaike information criterion (AIC) to assess the support for fluctuated directional or autocorrelated stabilizing selection on breeding time of the great tit population. Our results show that there is directional selection on the probability of breeding failure, and stabilizing selection on the mean number of fledglings. This selection for early laying date is consistent with a previous study of the same population. We also estimate the variation and autocorrelation in other parameters of the fitness functions, including the width and height, and found the height and location of annual fitness function are autocorrelated with significant variation, while the width can be assumed to constant over time. Using TMB to fit SSMs, we are able to estimate additional parameters compared to previous methods, all without requiring a substantial increase in computational resources. Furthermore, our specification of complex nonlinear model structure benefits greatly from the flexibility of model formulation with TMB. Therefore, our approach could be directly applied to estimating even more complicated phenotypic selection processes induced by environmental change for other species.

The school performance is a multidimensional phenomenon determined by the student, the school and the socio-economical context in which they are immerse. In order to measure the performance of the educational institutions, the Ministry of Education applies a set of standardized tests which form the Measurement System of Quality in Education, SIMCE for its initials in Spanish. There are some factors which determine the performance in the SIMCE. Along this line, many authors hold that the socio-economical level of the students is the most important one to explain the school performance. However, the data indicates that, not only the dependence of the educational institution but also the geographical location where it is located affect the school performance. This research aims to determine which are the best predictors in the SIMCE score of mathematics in educational institutions of the Araucania region. The nested data will be modelling by using a Generalized Linear Mixed Mode., El rendimiento académico es un fenómeno multidimensional, determinado por el estudiante, la escuela y el contexto socio-económico-cultural en que estos agentes se desenvuelven. Con el objeto de medir el rendimiento de los establecimientos educacionales, el Ministerio de Educación aplica una batería de pruebas estan darizadas que componen el Sistema de Medición de la Calidad de la Educación, SIMCE. Existen varios factores que determinan el rendimiento en el SIMCE. En este sentido, diversos autores sostienen que el nivel socio-económico de los estudiantes es el más importante para explicar el rendimiento escolar. Por otra parte, los datos indican que tanto la dependencia escolar como la zona geográfica donde está ubicado el establecimiento educacional influyen en el rendimiento. Este estudio tiene por objetivo determinar cu ́ales son los mejores predictores del puntaje SIMCE de matemáticas en establecimientos educacionales de la región de la Araucanía. Los datos de carácter anidado serían modelados a través de un Modelo Lineal Generalizado Mixto (GLMM).

With advances in whole-genome sequencing (WGS) technology, more advanced statistical methods for testing genetic association with rare variants are being developed. Methods in which variants are grouped for analysis are also known as variant-set, gene-based, and aggregate unit tests. The burden test and sequence kernel association test (SKAT) are two widely used variant-set tests, which were originally developed for samples of unrelated individuals and later have been extended to family data with known pedigree structures. However, computationally efficient and powerful variant-set tests are needed to make analyses tractable in large-scale WGS studies with complex study samples. In this paper, we propose the variant-set mixed model association tests (SMMAT) for continuous and binary traits using the generalized linear mixed model framework. These tests can be applied to large-scale WGS studies involving samples with population structure and relatedness, such as in the National Heart, Lung, and Blood Institute's Trans-Omics for Precision Medicine (TOPMed) program. SMMATs share the same null model for different variant sets, and a virtue of this null model, which includes covariates only, is that it needs to be fit only once for all tests in each genome-wide analysis. Simulation studies show that all the proposed SMMATs correctly control type I error rates for both continuous and binary traits in the presence of population structure and relatedness. We also illustrate our tests in a real data example of analysis of plasma fibrinogen levels in the TOPMed program (n = 23,763), using the Analysis Commons, a cloud-based computing platform.

Logistic linear mixed models are widely used in experimental designs and genetic analyses of binary traits. Motivated by modern applications, we consider the case of many groups of random effects, where each group corresponds to a variance component. When the number of variance components is large, fitting a logistic linear mixed model is challenging. Thus, we develop two efficient and stable minorization-maximization (MM) algorithms for estimating variance components based on a Laplace approximation of the logistic model. One of these leads to a simple iterative soft-thresholding algorithm for variance component selection using the maximum penalized approximated likelihood. We demonstrate the variance component estimation and selection performance of our algorithms by means of simulation studies and an analysis of real data.

Standardized relative abundance indices for swordfish (Xiphias gladius Linnaeus, 1758) caught by the Spanish surface longline in the western Mediterranean Sea were estimated for the period 1988-2017. Standardized CPUEs were estimated through a General Linear Mixed Modeling (GLMM) approach under a negative binomial (NB) error distribution assumption. The main factors in the standardization analysis were fishing area and time of the year (quarter). The standardized index showed notable annual fluctuations without any definite trend for the period under study.

Temporal and spatial variation in phenotypic selection due to changing environmental conditions is of great interest to evolutionary biologists, but few existing methods estimating its magnitude take into account the temporal autocorrelation. We use state-space models (SSMs) to analyse phenotypic selection processes that cannot be observed directly and use Template Model builder (TMB), an R package for computing and maximizing the Laplace approximation of the marginal likelihood for SSM and other complex, nonlinear latent variable model via automatic differentiation. Using a long-term great tit (Parus major) dataset, we fit several SSMs and conduct model selection based on Akaike information criterion (AIC) to assess the support for fluctuated directional or autocorrelated stabilizing selection on breeding time of the great tit population. Our results show that there is directional selection on the probability of breeding failure, and stabilizing selection on the mean number of fledglings. This selection for early laying date is consistent with a previous study of the same population. We also estimate the variation and autocorrelation in other parameters of the fitness functions, including the width and height, and found the height and location of annual fitness function are autocorrelated with significant variation, while the width can be assumed to constant over time. Using TMB to fit SSMs, we are able to estimate additional parameters compared to previous methods, all without requiring a substantial increase in computational resources. Furthermore, our specification of complex nonlinear model structure benefits greatly from the flexibility of model formulation with TMB. Therefore, our approach could be directly applied to estimating even more complicated phenotypic selection processes induced by environmental change for other species.

The school performance is a multidimensional phenomenon determined by the student, the school and the socio-economical context in which they are immerse. In order to measure the performance of the educational institutions, the Ministry of Education applies a set of standardized tests which form the Measurement System of Quality in Education, SIMCE for its initials in Spanish. There are some factors which determine the performance in the SIMCE. Along this line, many authors hold that the socio-economical level of the students is the most important one to explain the school performance. However, the data indicates that, not only the dependence of the educational institution but also the geographical location where it is located affect the school performance. This research aims to determine which are the best predictors in the SIMCE score of mathematics in educational institutions of the Araucania region. The nested data will be modelling by using a Generalized Linear Mixed Mode., El rendimiento académico es un fenómeno multidimensional, determinado por el estudiante, la escuela y el contexto socio-económico-cultural en que estos agentes se desenvuelven. Con el objeto de medir el rendimiento de los establecimientos educacionales, el Ministerio de Educación aplica una batería de pruebas estan darizadas que componen el Sistema de Medición de la Calidad de la Educación, SIMCE. Existen varios factores que determinan el rendimiento en el SIMCE. En este sentido, diversos autores sostienen que el nivel socio-económico de los estudiantes es el más importante para explicar el rendimiento escolar. Por otra parte, los datos indican que tanto la dependencia escolar como la zona geográfica donde está ubicado el establecimiento educacional influyen en el rendimiento. Este estudio tiene por objetivo determinar cu ́ales son los mejores predictores del puntaje SIMCE de matemáticas en establecimientos educacionales de la región de la Araucanía. Los datos de carácter anidado serían modelados a través de un Modelo Lineal Generalizado Mixto (GLMM).

Serial correlation and overdispersion must be handled properly in analyses of time series of counts, and parameter-driven models combine an underlying latent process with a conditional log-linear Poisson model (given the latent process) for that purpose. Regression coefficients have direct interpretations, but likelihood inference is not straight-forward. We consider a two-step procedure for estimation: First regression parameters are estimated from the marginal distribution; second parameters concerning the latent process are estimated with composite likelihood methods, based on low-order simultaneous or conditional distributions. Confidence intervals are computed by bootstrap. Properties of estimators are examined and compared to other methods in three simulation studies, and the methods are applied to two datasets from the literature concerning hospital admission related to asthma and traffic deaths.

Serial correlation and overdispersion must be handled properly in analyses of time series of counts, and parameter-driven models combine an underlying latent process with a conditional log-linear Poisson model (given the latent process) for that purpose. Regression coefficients have direct interpretations, but likelihood inference is not straight-forward. We consider a two-step procedure for estimation: First regression parameters are estimated from the marginal distribution; second parameters concerning the latent process are estimated with composite likelihood methods, based on low-order simultaneous or conditional distributions. Confidence intervals are computed by bootstrap. Properties of estimators are examined and compared to other methods in three simulation studies, and the methods are applied to two datasets from the literature concerning hospital admission related to asthma and traffic deaths.

Selection and breeding can be used to fight transmission of infectious diseases in livestock. The prevalence in a population depends on the susceptibility and infectivity of the animals. Knowledge on the genetic background of those traits would facilitate efficient selection for lower disease prevalence. We investigated the genetic background of host susceptibility and infectivity for digital dermatitis (DD), an endemic infectious claw disease in dairy cattle, with a genome-wide association study (GWAS), using either a simple linear mixed model or a generalized linear mixed model based on epidemiological theory. In total, 1,513 Holstein-Friesian cows of 12 Dutch dairy farms were scored for DD infection status and class (M0 to M4.1) every 2 wk for 11 times; 1,401 of these cows were genotyped with a 75k SNP chip. We performed a GWAS with a linear mixed model on 10 host disease status traits, and with a generalized linear mixed model with a complementary log-log link function (GLMM) on the probability that a cow would get infected between 2 scorings. With the GLMM, we fitted SNP effects for host susceptibility and host infectivity, while taking the variation in exposure of the susceptible cow to infectious herd mates into account. With the linear model we detected 4 suggestive SNP (false discovery rate < 0.20), 2 for the fraction of observations a cow had an active lesion on chromosomes 1 and 14, one for the fraction of observations a cow had an M2 lesion on at least one claw on chromosome 1 (the same SNP as for the fraction of observations with an active lesion), and one for the fraction of observations a cow had an M4.1 lesion on at least one claw on chromosome 10. Heritability estimates ranged from 0.09 to 0.37. With the GLMM we did not detect significant nor suggestive SNP. The SNP effects on disease status analyzed with the linear model had a correlation coefficient of only 0.70 with SNP effects on susceptibility of the GLMM, indicating that both models capture pa

Often sensory discrimination tests are performed with replications for the assessors. In this paper, we suggest a new way of analyzing data from a discrimination study. The model suggested in this paper is a Thurstonian mixed model, in which the variation from the assessors is modelled as a random effect in a generalized linear mixed model. The setting is a multi-product discrimination study with a binary paired comparison. This model makes it possible to embed the analyses of products into one analysis rather than having to do an analysis for each product separately. In addition, it is possible to embed the model into the Thurstonian framework obtaining d-prime interpretations of the estimates. Furthermore, it is possible to extract information about the assessors, even across the products. More specifically, assessor specific d-prime estimates are obtained providing a way to get information about the panel. These estimates are interesting because they make it possible to investigate if the assessors are assessing in a specific way.

Since the 1960s, companies aim to evaluate future performance of the business management to predict the medium term, processes of gestation and installation of statements of financial vulnerability. The information contained in the financial statements of companies and the ability to analyze the evolution in time of financial ratios allows building models predicting risk of financial crisis. This paper presents a risk prediction model based on the information contained in the financial statements of companies with public offering on the Santiago Stock Exchange and the Stock Exchange of Buenos Aires (Argentina) in the 2000s. The financial crisis is characterized by an inability to meet payment obligations, obtaining excessive quantities of waste and in extreme situations like bankruptcy and subsequent liquidation of the company. Until a little over a year ago, most of the work done to quantify the impact of financial ratios in business crisis apply cross-sectional models, but the construction of models for panel data (longitudinal studies) is relevant given that incorporate the temporal dimension in the study. In particular, it has been demonstrated that the mixed logistic model considered unobserved heterogeneity exceeds the performance of standard logistic model. Both Argentina and Chile have recently applied mixed models with random coefficients to predict statements of financial vulnerability. The results indicate that in Chilean companies, the ratio of working capital accounts for the largest proportion of heterogeneity induced correlation present data, which justifies its inclusion as random coefficient, while in the Argentine market it is profitability rate. In addition, as fixed effects, indicators best predictor of the financial crisis they are profitability ratios, rotation and debt. It is concluded that significant ratios have discriminatory power and their behavior shows that are indicators for predicting crises., Desde la década de 1960, las empresas tienen como objetivo evaluar los resultados futuros del gerenciamiento empresarial para predecir, a mediano plazo, procesos de gestación e instalación de estados de vulnerabiidad financiera. La información contenida en los estados financieros de las empresas y la posibilidad de analizar la evolución en el tiempo de los ratios contables permiten construir modelos de predicción de riesgo de crisis financiera. En este trabajo, se construye un modelo de predicción de riesgo con base en la información contenida en los Estados Contables de las empresas con oferta pública en las Bolsas de Santiago de Chile y de Valores de Buenos Aires (Argentina) para la década del 2000.La crisis financiera se caracteriza por la incapacidad de cumplir con las obligaciones de pago, la obtención de magnitudes excesivas de pérdidas y por situaciones extremas como la quiebra y posterior liquidación de la empresa. Hasta hace un poco más de un año, la mayoría de los trabajos desarrollados para cuantificar la incidencia de ratios financieros en la crisis empresarial, aplicaron métodos de corte transversal, por lo que la construcción de modelos para datos longitudinales resulta pertinente, en tanto incorporan la dimensión temporal en el estudio. En particular, se ha demostrado que el modelo logístico mixto, que tiene en cuenta la heterogeneidad no observada supera ampliamente el desempeño del modelo logístico estándar.Tanto en Argentina como en Chile se han aplicado, recientemente, modelos mixtos con coeficientes aleatorios para predecir estados de vulnerabilidad financiera.Los resultados obtenidos indican que, en las empresas chilenas, el ratio del capital de trabajo explica la mayor proporción de la heterogeneidad inducida por la correlación que presentan los datos, lo que justifica su inclusión como coeficiente aleatorio, mientras que en el mercado argentino lo es el índice de rentabilidad. Además, como efectos fijos, los indicadores con mayor capacidad pre

Longitudinal data occur frequently in medical studies and covariates measured by error are typical features of such data. Generalized linear mixed models (GLMMs) are commonly used to analyse longitudinal data. It is typically assumed that the random effects covariance matrix is constant across the subject (and among subjects) in these models. In many situations, however, this correlation structure may differ among subjects and ignoring this heterogeneity can cause the biased estimates of model parameters. In this thesis, following Lee et al. (2012), we propose an approach to properly model the random effects covariance matrix based on covariates in the class of GLMMs where we also have covariates measured by error. The resulting parameters from this decomposition have a sensible interpretation and can easily be modelled without the concern of positive definiteness of the resulting estimator. The performance of the proposed approach is evaluated through simulation studies which show that the proposed method performs very well in terms biases and mean square errors as well as coverage rates. The proposed method is also analysed using a data from Manitoba Follow-up Study.

Longitudinal data occur frequently in medical studies and covariates measured by error are typical features of such data. Generalized linear mixed models (GLMMs) are commonly used to analyse longitudinal data. It is typically assumed that the random effects covariance matrix is constant across the subject (and among subjects) in these models. In many situations, however, this correlation structure may differ among subjects and ignoring this heterogeneity can cause the biased estimates of model parameters. In this thesis, following Lee et al. (2012), we propose an approach to properly model the random effects covariance matrix based on covariates in the class of GLMMs where we also have covariates measured by error. The resulting parameters from this decomposition have a sensible interpretation and can easily be modelled without the concern of positive definiteness of the resulting estimator. The performance of the proposed approach is evaluated through simulation studies which show that the proposed method performs very well in terms biases and mean square errors as well as coverage rates. The proposed method is also analysed using a data from Manitoba Follow-up Study.

Since the 1960s, companies aim to evaluate future performance of the business management to predict the medium term, processes of gestation and installation of statements of financial vulnerability. The information contained in the financial statements of companies and the ability to analyze the evolution in time of financial ratios allows building models predicting risk of financial crisis. This paper presents a risk prediction model based on the information contained in the financial statements of companies with public offering on the Santiago Stock Exchange and the Stock Exchange of Buenos Aires (Argentina) in the 2000s. The financial crisis is characterized by an inability to meet payment obligations, obtaining excessive quantities of waste and in extreme situations like bankruptcy and subsequent liquidation of the company. Until a little over a year ago, most of the work done to quantify the impact of financial ratios in business crisis apply cross-sectional models, but the construction of models for panel data (longitudinal studies) is relevant given that incorporate the temporal dimension in the study. In particular, it has been demonstrated that the mixed logistic model considered unobserved heterogeneity exceeds the performance of standard logistic model. Both Argentina and Chile have recently applied mixed models with random coefficients to predict statements of financial vulnerability. The results indicate that in Chilean companies, the ratio of working capital accounts for the largest proportion of heterogeneity induced correlation present data, which justifies its inclusion as random coefficient, while in the Argentine market it is profitability rate. In addition, as fixed effects, indicators best predictor of the financial crisis they are profitability ratios, rotation and debt. It is concluded that significant ratios have discriminatory power and their behavior shows that are indicators for predicting crises., Desde la década de 1960, las empresas tienen como objetivo evaluar los resultados futuros del gerenciamiento empresarial para predecir, a mediano plazo, procesos de gestación e instalación de estados de vulnerabiidad financiera. La información contenida en los estados financieros de las empresas y la posibilidad de analizar la evolución en el tiempo de los ratios contables permiten construir modelos de predicción de riesgo de crisis financiera. En este trabajo, se construye un modelo de predicción de riesgo con base en la información contenida en los Estados Contables de las empresas con oferta pública en las Bolsas de Santiago de Chile y de Valores de Buenos Aires (Argentina) para la década del 2000.La crisis financiera se caracteriza por la incapacidad de cumplir con las obligaciones de pago, la obtención de magnitudes excesivas de pérdidas y por situaciones extremas como la quiebra y posterior liquidación de la empresa. Hasta hace un poco más de un año, la mayoría de los trabajos desarrollados para cuantificar la incidencia de ratios financieros en la crisis empresarial, aplicaron métodos de corte transversal, por lo que la construcción de modelos para datos longitudinales resulta pertinente, en tanto incorporan la dimensión temporal en el estudio. En particular, se ha demostrado que el modelo logístico mixto, que tiene en cuenta la heterogeneidad no observada supera ampliamente el desempeño del modelo logístico estándar.Tanto en Argentina como en Chile se han aplicado, recientemente, modelos mixtos con coeficientes aleatorios para predecir estados de vulnerabilidad financiera.Los resultados obtenidos indican que, en las empresas chilenas, el ratio del capital de trabajo explica la mayor proporción de la heterogeneidad inducida por la correlación que presentan los datos, lo que justifica su inclusión como coeficiente aleatorio, mientras que en el mercado argentino lo es el índice de rentabilidad. Además, como efectos fijos, los indicadores con mayor capacidad pre

The purpose of this dissertation was to examine patterns of antenatal stressful life events (SLEs) experienced by women in the United States (U.S.) and their association with postpartum depression (PPD). It further explored the role of women's state-level socio-economic status (SES) on PPD; the racial/ethnic dispartites in SLE-PPD relationship; and the role of provider communication on perinatal depression. Data from 2009–11 Pregnancy Risk Assessment Monitoring System (PRAMS) and SES indicators published by the Institute of Women’s Policy Research (IWPR) were used. Latent class analysis (LCA) was performed to identify unobserved class membership based on antenatal SLEs. Multilevel generalized linear mixed models examined whether state-level SES moderated the antenatal SLE-PPD relationship. Of 116,595 respondents to the PRAMS 2009-11, the sample size for our analyses ranged from 78% to 99%. The majority (64%) of participants were in low-stress class. The illness/death related-stress class (13%) had a high prevalence of severe illness (77%) and death (63%) of a family member or someone very close to them, while those in the multiple-stress (22%) class endorsed most other SLEs. Eleven percent had PPD; women who experienced all types of stressors, had the highest odds (adjusted odds ratio [aOR]: 5.43; 95% confidence interval [CI]: 5.36, 5.51) of PPD. The odds of PPD decreased with increasing state-level social/economic autonomy index (aOR: 0.75; 95% CI: 0.64, 0.88), with significant cross-level interaction between stressors and state-level SES. Among non-Hispanic blacks and non-Hispanic whites, husband/partner not wanting the pregnancy (aOR: 1.47; 95% CI: 1.14, 1.90) and drug/drinking problems of someone close (aOR: 1.37; 95% CI: 1.21, 1.55) were respectively associated with PPD. Provider communication was protective. That 1 out of every 5 and 1 out of every 8 women were in the high- and emotional-stress classes suggests that SLEs are common among pregnant women. Our re

Longitudinal data occur frequently in medical studies and covariates measured by error are typical features of such data. Generalized linear mixed models (GLMMs) are commonly used to analyse longitudinal data. It is typically assumed that the random effects covariance matrix is constant across the subject (and among subjects) in these models. In many situations, however, this correlation structure may differ among subjects and ignoring this heterogeneity can cause the biased estimates of model parameters. In this thesis, following Lee et al. (2012), we propose an approach to properly model the random effects covariance matrix based on covariates in the class of GLMMs where we also have covariates measured by error. The resulting parameters from this decomposition have a sensible interpretation and can easily be modelled without the concern of positive definiteness of the resulting estimator. The performance of the proposed approach is evaluated through simulation studies which show that the proposed method performs very well in terms biases and mean square errors as well as coverage rates. The proposed method is also analysed using a data from Manitoba Follow-up Study.

Longitudinal data occur frequently in medical studies and covariates measured by error are typical features of such data. Generalized linear mixed models (GLMMs) are commonly used to analyse longitudinal data. It is typically assumed that the random effects covariance matrix is constant across the subject (and among subjects) in these models. In many situations, however, this correlation structure may differ among subjects and ignoring this heterogeneity can cause the biased estimates of model parameters. In this thesis, following Lee et al. (2012), we propose an approach to properly model the random effects covariance matrix based on covariates in the class of GLMMs where we also have covariates measured by error. The resulting parameters from this decomposition have a sensible interpretation and can easily be modelled without the concern of positive definiteness of the resulting estimator. The performance of the proposed approach is evaluated through simulation studies which show that the proposed method performs very well in terms biases and mean square errors as well as coverage rates. The proposed method is also analysed using a data from Manitoba Follow-up Study.

Understanding the spatial and temporal scales at which environmental variation affects populations of plants and animals is an important goal for modern population biology, especially in the context of shifting climatic conditions. The El Niño Southern Oscillation (ENSO) generates climatic extremes of interannual variation, and has been shown to have significant effects on the diversity and abundance of a variety of terrestrial taxa. However, studies that have investigated the influence of such large-scale climate phenomena have often been limited in spatial and taxonomic scope. We used 23 years (1988-2010) of a long-term butterfly monitoring data set to explore associations between variation in population abundance of 28 butterfly species and variation in ENSO-derived sea surface temperature anomalies (SSTA) across 10 sites that encompass an elevational range of 2750 m in the Sierra Nevada mountain range of California. Our analysis detected a positive, regional effect of increased SSTA on butterfly abundance (wetter and warmer years predict more butterfly observations), yet the influence of SSTA on butterfly abundances varied along the elevational gradient, and also differed greatly among the 28 species. Migratory species had the strongest relationships with ENSO-derived SSTA, suggesting that large-scale climate indices are particularly valuable for understanding biotic-abiotic relationships of the most mobile species. In general, however, the ecological effects of large-scale climatic factors are context dependent between sites and species. Our results illustrate the power of long-term data sets for revealing pervasive yet subtle climatic effects, but also caution against expectations derived from exemplar species or single locations in the study of biotic-abiotic interactions.

This dissertation aims to develop statistical methodologies for estimating the effects of time-fixed and time-varying factors in recurrent events modeling context. The research is motivated by the traffic safety research question of evaluating the influence of crash on driving risk and driver behavior. The methodologies developed, however, are general and can be applied to other fields. Four alternative approaches based on various data settings are elaborated and applied to 100-Car Naturalistic Driving Study in the following Chapters. Chapter 1 provides a general introduction and background of each method, with a sketch of 100-Car Naturalistic Driving Study. In Chapter 2, I assessed the impact of crash on driving behavior by comparing the frequency of distraction events in per-defined windows. A count-based approach based on mixed-effect binomial regression models was used. In Chapter 3, I introduced intensity-based recurrent event models by treating number of Safety Critical Incidents and Near Crash over time as a counting process. Recurrent event models fit the natural generation scheme of the data in this study. Four semi-parametric models are explored: Andersen-Gill model, Andersen-Gill model with stratified baseline functions, frailty model, and frailty model with stratified baseline functions. I derived model estimation procedure and and conducted model comparison via simulation and application. The recurrent event models in Chapter 3 are all based on proportional assumption, where effects are constant. However, the change of effects over time is often of primary interest. In Chapter 4, I developed time-varying coefficient model using penalized B-spline function to approximate varying coefficients. Shared frailty terms was used to incorporate correlation within subjects. Inference and statistical test are also provided. Frailty representation was proposed to link time-varying coefficient model with regular frailty model. In Chapter 5, I further extended framew

Understanding the spatial and temporal scales at which environmental variation affects populations of plants and animals is an important goal for modern population biology, especially in the context of shifting climatic conditions. The El Niño Southern Oscillation (ENSO) generates climatic extremes of interannual variation, and has been shown to have significant effects on the diversity and abundance of a variety of terrestrial taxa. However, studies that have investigated the influence of such large-scale climate phenomena have often been limited in spatial and taxonomic scope. We used 23 years (1988-2010) of a long-term butterfly monitoring data set to explore associations between variation in population abundance of 28 butterfly species and variation in ENSO-derived sea surface temperature anomalies (SSTA) across 10 sites that encompass an elevational range of 2750 m in the Sierra Nevada mountain range of California. Our analysis detected a positive, regional effect of increased SSTA on butterfly abundance (wetter and warmer years predict more butterfly observations), yet the influence of SSTA on butterfly abundances varied along the elevational gradient, and also differed greatly among the 28 species. Migratory species had the strongest relationships with ENSO-derived SSTA, suggesting that large-scale climate indices are particularly valuable for understanding biotic-abiotic relationships of the most mobile species. In general, however, the ecological effects of large-scale climatic factors are context dependent between sites and species. Our results illustrate the power of long-term data sets for revealing pervasive yet subtle climatic effects, but also caution against expectations derived from exemplar species or single locations in the study of biotic-abiotic interactions.

When performing single arm meta-analyses of rare events in small populations, if the outcome of interest is incidence, it is not uncommon to have at least one study with zero events, especially in the presence of competing risks. In this paper, we address the problem of how to include studies with zero events in inverse variance meta-analyses when individual patient data are not available, going beyond the naïve approach of not including the study or the use of a continuity correction. The proposed solution is the arcsine transformation of the crude cumulative incidence as its approximate variance, which is inversely proportional to the sample size, can be calculated also for studies with a zero estimate. As an alternative, generalized linear mixed models (GLMM) can be used. Simulations were performed to compare the results from inverse variance method meta-analyses of the arcsine transformed cumulative incidence to those obtained from meta-analyses of the cumulative incidence itself and of the logit transformation of the cumulative incidence. The comparisons have been carried out for different scenarios of heterogeneity, incidence, and censoring and for competing and not competing risks. The arcsine transformation showed the smallest bias and the highest coverage among models assuming within study normality. At the same time, the GLMM model had the best performance at very low incidences. The proposed method was applied to the clinical context that motivated this work, i.e. a meta-analysis of 5-year crude cumulative incidence of central nervous system recurrences in children treated for acute lymphoblastic leukemia.

Providing quantitative management guidelines is essential for an effective conservation of forest-dependent animal communities. Traditional forest practices at the stand scale simultaneously alter both physical and floristic features with a negative effect on ecosystem processes. Thus, we tested and proposed a method to define forestry prescriptions taking into account the combined effect of woodland structure and tree species composition on the presence of four bird indicator species (Marsh Tit Poecile palustris, European Nuthatch Sitta europaea, Short-toed Tree-creeper Certhya brachydactyla and Blue Tit Cyanistes caeruleus). The study was carried out in Lombardy (Northern Italy), from 2002 to 2005. By using a stratified cluster sampling design, we recorded Basal Area, one hundred tree trunk diameters at breast height (DBH) and tree species in 160 sampling plots, grouped in 23 sampling areas. In each plot we also performed a bird survey using the point count method. We analyzed data using Multimodel Inference and Model Averaging on Generalized Linear Mixed Models, with species presence/absence as the response variable, sampling area as a random factor and forest covariates as fixed factors. In order to test our method, we compared it with other two traditional approaches, which consider structural and tree floristic variables separately. Model comparison showed that our method performed better than traditional ones, in both the evaluation and validation processes. Based on our main results, in deciduous mixed forest where the exploitation demand is limited, we recommend maintaining at least 65 trees/ha with DBH > 45 cm. In particular, we advise keeping 70 trees/ha with DBH > 50 cm in chestnut forests and 300 trees/ha with DBH 20-30 cm in oak forests. Conversely, in more exploited oak forests, we advise maintaining at least 670 trees/ha with DBH 15-30 cm in chestnut forests and 100 trees/ha with DBH 10-15 cm.

The Intraclass Correlation Coefficient (ICC) is commonly used to estimate the similarity between quantitative measures obtained from different sources. Overdispersed data is traditionally transformed so that linear mixed model (LMM) based ICC can be estimated. A common transformation used is the natural logarithm. The reliability of environmental sampling of fecal slurry on freestall pens has been estimated for Mycobacterium avium subsp. paratuberculosis using the natural logarithm transformed culture results. Recently, the negative binomial ICC was defined based on a generalized linear mixed model for negative binomial distributed data. The current study reports on the negative binomial ICC estimate which includes fixed effects using culture results of environmental samples. Simulations using a wide variety of inputs and negative binomial distribution parameters (r; p) showed better performance of the new negative binomial ICC compared to the ICC based on LMM even when negative binomial data was logarithm, and square root transformed. A second comparison that targeted a wider range of ICC values showed that the mean of estimated ICC closely approximated the true ICC.

The Intraclass Correlation Coefficient (ICC) is commonly used to estimate the similarity between quantitative measures obtained from different sources. Overdispersed data is traditionally transformed so that linear mixed model (LMM) based ICC can be estimated. A common transformation used is the natural logarithm. The reliability of environmental sampling of fecal slurry on freestall pens has been estimated for Mycobacterium avium subsp. paratuberculosis using the natural logarithm transformed culture results. Recently, the negative binomial ICC was defined based on a generalized linear mixed model for negative binomial distributed data. The current study reports on the negative binomial ICC estimate which includes fixed effects using culture results of environmental samples. Simulations using a wide variety of inputs and negative binomial distribution parameters (r; p) showed better performance of the new negative binomial ICC compared to the ICC based on LMM even when negative binomial data was logarithm, and square root transformed. A second comparison that targeted a wider range of ICC values showed that the mean of estimated ICC closely approximated the true ICC.

Religious rituals that are painful or highly stressful are hypothesized to be costly signs of commitment essential for the evolution of complex society. Yet few studies have investigated how such extreme ritual practices were culturally transmitted in past societies. Here, we report the first study to analyze temporal and spatial variation in bloodletting rituals recorded in Classic Maya (ca. 250-900 CE) hieroglyphic texts. We also identify the sociopolitical contexts most closely associated with these ancient recorded rituals. Sampling an extensive record of 2,480 hieroglyphic texts, this study identifies every recorded instance of the logographic sign for the word ch'ahb' that is associated with ritual bloodletting. We show that documented rituals exhibit low frequency whose occurrence cannot be predicted by spatial location. Conversely, network ties better capture the distribution of bloodletting rituals across the southern Maya region. Our results indicate that bloodletting rituals by Maya nobles were not uniformly recorded, but were typically documented in association with antagonistic statements and may have signaled royal commitments among connected polities.

1.The ambrosia beetle Platypus quercivorus uses fagaceous tree species as its hosts, and causes Japanese oak wilt by transporting and introducing the pathogenic fungus Raffaelea quercivora into the host tree. To protect fagaceous trees, it is necessary to understand the host selection process used by the beetle in heterogeneous forests. The occurrence of beetle attack on Quercus crispula and Castanea crenata was monitored from 2008 to 2011 in two 1-ha and one 0.5-ha plots established in a secondary forest. The stem and crown densities of each fagaceous tree species were calculated at 11 different spatial scales for each individual tree. Model application and selection were performed to clarify factors affecting the probability of attack by P. quercivorus. As shown in previous studies, a higher probability of attack was predicted for Q. crispula than for C. crenata, and the predicted probability of attack was lower for trees attacked in the previous year. The effect of stem density on the probability of attack was positive at a small (radius 5 m) spatial scale and negative at larger (radius 5–17.5 m) scales. We interpreted this to indicate that a dense distribution of fagaceous trees in a small area had a higher probability of attack, and that this probability decreased with an increasing density of fagaceous trees in a surrounding concentric area. A positive effect of crown density was detected, suggesting that P. quercivorus has a means of detecting host trees in areas with high crown densities of fagaceous species.

The Spanish economic crisis has led to a significant reduction in housing sales, and therefore, there has been a decrease in housing prices. In this paper, we analyze changes in the average housing price throughout Spain. We use quarterly data from a random sample of 150 municipalities from the first quarter (Q1) of 2005, before the financial crisis started, to Q1 2010. Our analysis uses generalized estimating equation and generalized linear mixed model approaches. Data published for Q2, Q3, and Q4 2010 are compared with the data fitted using these models. Finally, the methods are compared with time-series models., Depto. de Estadística e Investigación Operativa, Fac. de Ciencias Matemáticas, TRUE, pub

Motivated by the recent discourses on approaches to combine generalizability theory (GT) and item response theory (IRT), this study suggests an approach that answers to some of the issues raised by previous research on combining GT and IRT. The main idea of the proposed approach is to recognize that IRT models can be written in terms of a latent continuous response and that a classic IRT model can be modeled directly using standard GT with items as a fixed facet. Once this is recognized, treating items as a random facet or considering other facets such as raters, become relatively straightforward extensions. The resulting models are logistic mixed models that contain the parameter of interest: the variance components needed for the generalizability and dependability coefficients. The models can be estimated by straightforward ML estimation. Extensive simulation studies were conducted to evaluate the performance of the proposed approach under various measurement situations. The results suggested that the proposed approach gives overall accurate results, in particular, when estimating the generalizability coefficients. The use of the proposed method is illustrated with large-scale data sets from classroom assessment (the Carbon Cycle 2008-2009 data set) and standardized tests (the PISA 2009 U.S. Science data set). The empirical results were presented and interpreted in the context of the G study and the D study in GT as well as a Wright Map in IRT. The results demonstrated the flexibility of the proposed approach with respect to incorporating extra complications in measurement situations (e.g., multidimensionality, polytomous responses) and further explanatory variables (e.g., rater facet).

In modern Item Response Theory, the Rasch model is viewed as a Generalized Linear Mixed Model, where the item parameters correspond to the fixed-effects, whereas the person specific parameters are the random-effects. The statistical model, bearing on the observable variables only, is obtained after integrating out the random-effects. Although it is widely accepted that the parameters of this model are identified, it is hard to find a correct justification. Furthermore, the meaning of the parameters of the Rasch model – as well as of its extensions – is typically based on the fixed-effects specification of the model, that is, when the person specific parameters are also treated as fixed-effects. The contribution of this paper is to provide an explicit proof of the identification of the random-effects Rasch model. The proof is valid for a large class of Rasch-type models. It is also shown that such a proof can be applied to analyze the identification of Explanatory Rasch Models. Finally, the meaning of the parameters of interest with respect to the different data generating process is discussed.

Motivated by the recent discourses on approaches to combine generalizability theory (GT) and item response theory (IRT), this study suggests an approach that answers to some of the issues raised by previous research on combining GT and IRT. The main idea of the proposed approach is to recognize that IRT models can be written in terms of a latent continuous response and that a classic IRT model can be modeled directly using standard GT with items as a fixed facet. Once this is recognized, treating items as a random facet or considering other facets such as raters, become relatively straightforward extensions. The resulting models are logistic mixed models that contain the parameter of interest: the variance components needed for the generalizability and dependability coefficients. The models can be estimated by straightforward ML estimation. Extensive simulation studies were conducted to evaluate the performance of the proposed approach under various measurement situations. The results suggested that the proposed approach gives overall accurate results, in particular, when estimating the generalizability coefficients. The use of the proposed method is illustrated with large-scale data sets from classroom assessment (the Carbon Cycle 2008-2009 data set) and standardized tests (the PISA 2009 U.S. Science data set). The empirical results were presented and interpreted in the context of the G study and the D study in GT as well as a Wright Map in IRT. The results demonstrated the flexibility of the proposed approach with respect to incorporating extra complications in measurement situations (e.g., multidimensionality, polytomous responses) and further explanatory variables (e.g., rater facet).

Niche theory predicts that species occurrence and abundance show non-linear, unimodal relationships with respect to environmental gradients. Unimodal models, such as the Gaussian (logistic) model, are however more difficult to fit to data than linear ones, particularly in a multi-species context in ordination, with trait modulated response and when species phylogeny and species traits must be taken into account. Adding squared terms to a linear model is a possibility but gives uninterpretable parameters. This paper explains why and when generalized linear mixed models, even without squared terms, can effectively analyse unimodal data and also presents a graphical tool and statistical test to test for unimodal response while fitting just the generalized linear mixed model. The R-code for this is supplied in Supplemental Information 1

Various multi-label classification algorithms are broadly developed in the literature. Each of these existing methods has different strengths and drawbacks, and very few attempts have been performed to address the uncertainty. Recently, neutral zone classifiers have been presented to deal with ambiguous data to improve the accuracy of the estimated classification results by trading off the relatively high penalty of an incorrect classification with a lower penalty for remaining neutral about class membership. The existing neutral zone methods was developed to construct multiple single-label classifiers for multi-label data. In this work, we present different classification alternatives for multi-label data using a standard logistic regression model, a generalized linear mixed model and Markov random field to relax the current underlying assumptions. The proposed neutral zone classification methods are implemented and tested on simulation data sets and on the biological data set. Their results suggest that our proposed classification approaches are useful alternatives for practical application when working with multi-label data.

Various multi-label classification algorithms are broadly developed in the literature. Each of these existing methods has different strengths and drawbacks, and very few attempts have been performed to address the uncertainty. Recently, neutral zone classifiers have been presented to deal with ambiguous data to improve the accuracy of the estimated classification results by trading off the relatively high penalty of an incorrect classification with a lower penalty for remaining neutral about class membership. The existing neutral zone methods was developed to construct multiple single-label classifiers for multi-label data. In this work, we present different classification alternatives for multi-label data using a standard logistic regression model, a generalized linear mixed model and Markov random field to relax the current underlying assumptions. The proposed neutral zone classification methods are implemented and tested on simulation data sets and on the biological data set. Their results suggest that our proposed classification approaches are useful alternatives for practical application when working with multi-label data.

In this study, I investigated how proficiently seventh-grade students enrolled in two Southwestern schools solve algebra word problems. I analyzed various factors that could affect this proficiency and explored the differences between English Learners (ELs) and native English Primary students (EPs). I collected the data as part of the Animal Watch project, a computer-based initiative designed to improve the mathematical skills of children from grades 5-8 in the Southwest. A sample of 86 students (26 ELs and 60 EPs), clustered in four different classes, was used for this project. A Generalized Linear Mixed Model (GLMM) approach with the GLIMMIX procedure in SAS 9.3 showed that students from the classes that had a higher percentage of EL students performed better than those in the classes where the EL concentration was lower. Classes with more EL males were better at learning mathematics than classes with more EP females. The results also indicated: (a) a positive correlation between the students' ability to solve algebra word problems on their first attempt and their success ratio in solving all problems, and (b) a negative correlation between the percentage of problems solved correctly and those considered too hard from the very beginning. I conclude my dissertation by making specific recommendations for further research.

Generalized linear mixed models or latent variable models for categorical data are difficult to estimate if the random effects or latent variables vary at non-nested levels, such as persons and test items. Clayton and Rasbash (1999) suggested an Alternating Imputation Posterior (AIP) algorithm for approximate maximum likelihood estimation. For item response models with random item effects, the algorithm iterates between an item wing in which the item mean and variance are estimated for given person effects and a person wing in which the person mean and variance are estimated for given item effects. The person effects used for the item wing are sampled from the conditional posterior distribution estimated in the person wing and vice versa. Clayton and Rasbash (1999) used marginal quasi-likelihood (MQL) and penalized quasi-likelihood (PQL) estimation within the AIP algorithm, but this method has been shown to produce biased estimates in many situations, so we use maximum likelihood estimation with adaptive quadrature. We apply the proposed algorithm to the famous salamander mating data, comparing the estimates with many other methods, and to an educational testing dataset. We also present a simulation study to assess performance of the AIP algorithm and the Laplace approximation with different numbers of items and persons and a range of item and person variances. © 2010 Elsevier B.V. All rights reserved.

With growing population and fast urbanization in Australia, it is a challenging task to maintain our water quality. It is essential to develop an appropriate statistical methodology in analyzing water quality data in order to draw valid conclusions and hence provide useful advices in water management. This paper is to develop robust rank-based procedures for analyzing nonnormally distributed data collected over time at different sites. To take account of temporal correlations of the observations within sites, we consider the optimally combined estimating functions proposed by Wang and Zhu (Biometrika, 93:459-464, 2006) which leads to more efficient parameter estimation. Furthermore, we apply the induced smoothing method to reduce the computational burden. Smoothing leads to easy calculation of the parameter estimates and their variance-covariance matrix. Analysis of water quality data from Total Iron and Total Cyanophytes shows the differences between the traditional generalized linear mixed models and rank regression models. Our analysis also demonstrates the advantages of the rank regression models for analyzing nonnormal data.

Generalized linear mixed models or latent variable models for categorical data are difficult to estimate if the random effects or latent variables vary at non-nested levels, such as persons and test items. Clayton and Rasbash (1999) suggested an Alternating Imputation Posterior (AIP) algorithm for approximate maximum likelihood estimation. For item response models with random item effects, the algorithm iterates between an item wing in which the item mean and variance are estimated for given person effects and a person wing in which the person mean and variance are estimated for given item effects. The person effects used for the item wing are sampled from the conditional posterior distribution estimated in the person wing and vice versa. Clayton and Rasbash (1999) used marginal quasi-likelihood (MQL) and penalized quasi-likelihood (PQL) estimation within the AIP algorithm, but this method has been shown to produce biased estimates in many situations, so we use maximum likelihood estimation with adaptive quadrature. We apply the proposed algorithm to the famous salamander mating data, comparing the estimates with many other methods, and to an educational testing dataset. We also present a simulation study to assess performance of the AIP algorithm and the Laplace approximation with different numbers of items and persons and a range of item and person variances. © 2010 Elsevier B.V. All rights reserved.

The generalized linear model (GLM) is a popular model in many research areas. In the GLM, each outcome of the dependent variable is assumed to be generated from a particular distribution function in the exponential family. The mean of the distribution depends on the independent variables. The link function provides the relationship between the linear predictor and the mean of the distribution function. In this dissertation, two semiparametric extensions of the GLM will be developed. In the first part of this dissertation, we have proposed a new model, called a semiparametric generalized linear model with a log-concave random component (SGLM-L). In this model, the estimate of the distribution of the random component has a nonparametric form while the estimate of the systematic part has a parametric form. In the second part of this dissertation, we have proposed a model, called a generalized semiparametric single-index mixed model (GSSIMM). A nonparametric component with a single index is incorporated into the mean function in the generalized linear mixed model (GLMM) assuming that the random component is following a parametric distribution. In the first part of this dissertation, since most of the literature on the GLM deals with the parametric random component, we relax the parametric distribution assumption for the random component of the GLM and impose a log-concave constraint on the distribution. An iterative numerical algorithm for computing the estimators in the SGLM-L is developed. We construct a log-likelihood ratio test for inference. In the second part of this dissertation, we use a single index model to generalize the GLMM to have a linear combination of covariates enter the model via a nonparametric mean function, because the linear model in the GLMM is not complex enough to capture the underlying relationship between the response and its associated covariates. The marginal likelihood is approximated using the Laplace method. A penalized quasi-likelihood