Experimental Design and Statistical Inference: Generalized Least Squares and Repeated Measures over Time

2015 ◽  
pp. 85-96
Author(s):  
F. G. Giesbrecht
Keyword(s):  
Experimental Design ◽  
Least Squares ◽  
Statistical Inference ◽  
Repeated Measures ◽  
Generalized Least Squares ◽  
Over Time

Spatiotemporal palaeodiversity patterns of modern crocodiles (Crocodyliformes: Eusuchia)

2019 ◽  
Vol 189 (2) ◽  
pp. 635-656 ◽  
Author(s):  
Ane De Celis ◽  
Iván Narváez ◽  
Francisco Ortega
Keyword(s):  
South America ◽  
Least Squares ◽  
Late Miocene ◽  
North American ◽  
Fossil Record ◽  
Generalized Least Squares ◽  
Global Scale ◽  
Biotic Factors ◽  
Species Dynamics ◽  
Over Time

Abstract Eusuchia is a crocodyliform clade with a rich and diverse fossil record dating back to the Mesozoic. There are several recent studies that analyse crocodyliform palaeodiversity over time, but none of them focuses exclusively on eusuchians. Thus, we estimated subsampled eusuchian palaeodiversity species dynamics over time not only at a global scale, but also by continents and main crocodylian lineages (Alligatoroidea, Crocodyloidea and Gavialoidea). These estimates reveal complex spatiotemporal palaeodiversity patterns, in which two maxima can be detected: the first during the Palaeocene and the second, which is also the biggest, in the middle-late Miocene. The Palaeocene shift is related to a North American alligatoroid diversification, whereas the middle–late Miocene maximum is related to a diversification of the three main Crocodylia lineages in Gondwanan land masses, but especially in South America. Additionally, a model-based study using generalized least squares was carried out to analyse the relationships between different abiotic and sampling proxies and eusuchian palaeodiversity. The results show that palaeotemperature is the most important factor amongst the analysed proxies, in accordance with previous studies. However, the results suggest that, along with palaeotemperature, other abiotic and/or biotic factors might also be driving eusuchian palaeodiversity dynamics.

Estimating an Autoregressive Current Effects Model of Sales Response when Observations are Aggregated over Time: Least Squares versus Maximum Likelihood

1988 ◽  
Vol 25 (3) ◽  
pp. 301-307
Author(s):  
Wilfried R. Vanhonacker
Keyword(s):  
Maximum Likelihood ◽  
Least Squares ◽  
Serial Correlation ◽  
Mean Squared Error ◽  
Generalized Least Squares ◽  
Parameter Estimates ◽  
Squared Error ◽  
Positive Serial Correlation ◽  
Serial Correlation Coefficient ◽  
Over Time

Estimating autoregressive current effects models is not straightforward when observations are aggregated over time. The author evaluates a familiar iterative generalized least squares (IGLS) approach and contrasts it to a maximum likelihood (ML) approach. Analytic and numerical results suggest that (1) IGLS and ML provide good estimates for the response parameters in instances of positive serial correlation, (2) ML provides superior (in mean squared error) estimates for the serial correlation coefficient, and (3) IGLS might have difficulty in deriving parameter estimates in instances of negative serial correlation.

Power Differences Between the Modified Brown-Forsythe and Mixed-Model Approaches in Repeated Measures Designs

Methodology ◽  
2007 ◽  
Vol 3 (1) ◽  
pp. 1-13 ◽  
Author(s):  
Guillermo Vallejo ◽  
María P. Fernández ◽  
Pablo E. Livacic-Rojas
Keyword(s):  
Least Squares ◽  
Covariance Matrix ◽  
Interaction Effect ◽  
Repeated Measures ◽  
Mixed Model ◽  
Information Criterion ◽  
Generalized Least Squares ◽  
Superior Performance ◽  
Multivariate Normality ◽  
Repeated Measures Designs

Abstract. This article compares the sensibility of the modified Brown-Forsythe (MBF) approach developed by Vallejo and Ato (2006) and a modified empirical generalized least squares (EGLS) method adjusted by the Kenward-Roger solution available in the SAS Institute's (2002) Proc Mixed program to detect the presence of an interaction effect under departures from covariance homogeneity and multivariate normality. Although none of the approaches demonstrated superior performance in all situations, our results indicate that the so-called EGLS method, based on the Akaike's Information Criterion or based on always assuming a unstructured between-subjects heterogeneous covariance pattern, was the most powerful alternative. Results also indicate that little power can be achieved with the EGLS method if the covariance matrix is specified correctly.

scholarly journals MODELAGEM NÃO LINEAR DO CRESCIMENTO EM ALTURA DO CAFEEIRO IRRIGADO E NÃO IRRIGADO EM DIFERENTES DENSIDADES

Irriga ◽  
2018 ◽  
Vol 1 (1) ◽  
pp. 140
Author(s):  
Adriele Aparecida Pereira ◽  
Tales Jesus Fernandes ◽  
Myriane Stella Scalco ◽  
Augusto Ramalho De Morais
Keyword(s):  
Least Squares ◽  
Least Squares Method ◽  
Gompertz Model ◽  
Generalized Least Squares ◽  
Residual Variance ◽  
Growth Data ◽  
Nonlinear Growth ◽  
Heterogeneity Of Variance ◽  
E Mail ◽  
Over Time

MODELAGEM NÃO LINEAR DO CRESCIMENTO EM ALTURA DO CAFEEIRO IRRIGADO E NÃO IRRIGADO EM DIFERENTES DENSIDADES  ADRIELE APARECIDA PEREIRA1; TALES JESUS FERNANDES2; MYRIANE STELLA SCALCO3 E AUGUSTO RAMALHO DE MORAIS4 1Licenciada em Matemática, Mestre, DEX/UFLA, Lavras-MG, e-mail: [email protected] em Matemática, Doutor, Prof. DEX/UFLA, Lavras-MG, e-mail: [email protected] Agrônoma, Doutora, DAG/UFLA, Lavras-MG, e-mail: [email protected] Agrônomo, Doutor, Prof. DEX/UFLA, Lavras-MG, e-mail: [email protected]  1 RESUMO Heterogeneidade de variâncias e autocorrelação residual são características inerentes à dados de crescimento ao longo do tempo que se não considerados nas análises podem conduzir a resultados imprecisos. Este estudo teve por objetivo comparar os ajustes dos modelos Logístico e Gompertz, considerando os métodos de mínimos quadrados: ordinários e generalizados. Os dados utilizados referem-se à altura de plantas do cafeeiro, submetidas aos regimes de irrigação Si (testemunha), 60 kPa e 140 kPa, nas densidades de plantio 2500 e 5000 plantas ha-1. Segundo o desvio padrão residual e a análise de resíduos, o ajuste do modelo Gompertz pelo método de mínimos quadrados generalizados, que incorpora a heterogeneidade de variâncias e autocorrelação residual na modelagem, apresentou os melhores resultados para todos os dados analisados, sendo indicado para modelar o crescimento em altura do cafeeiro ao longo do tempo. Os ajustes referentes às plantas irrigadas apresentaram as maiores estimativas para a altura assintótica, confirmando que a irrigação da lavoura proporciona maior crescimento das plantas. Palavras-Chave: Autocorrelação residual, Gompertz, Heterocedasticidade.  PEREIRA, A. A.; FERNANDES, T. J.; SCALCO, M. S.; MORAIS, A. R. de MODELING NONLINEAR GROWTH IN HEIGHT COFFEE WITH AND WITHOUT IRRIGATION IN DIFFERENT DENSITIES  2 ABSTRACT Heterogeneity of variance and residual autocorrelation characteristics are inherent in the growth data over time that is not considered in the analysis may lead to inaccurate results. This study aimed to compare the settings of the Logistic and Gompertz models, considering the methods of least squares: ordinary and generalized. The data used refer to the height of the coffee plants, subjected to irrigation systems Si (non irrigated), 60 kPa and 140 kPa, the planting densities in 2500 and 5000 plants ha-1. According to the residual standard deviation and the residual analysis, the fit of the Gompertz model by generalized least squares method, which incorporates the heterogeneity of residual variance and autocorrelation in modeling, showed the best results for all data analyzed, suitable for modeling the growth in height of the coffee over time. The adjustments related to the irrigated plants had the highest estimates for the asymptotic height, confirming that the crop irrigation provides greater plant growth. Keywords: Residual autocorrelation, Gompertz, Heteroscedasticity.

Improving plant biomass estimation in the field using partial least squares regression and ridge regression

Botany ◽  
2016 ◽  
Vol 94 (7) ◽  
pp. 501-508 ◽  
Author(s):  
Brian M. Ohsowski ◽  
Kari E. Dunfield ◽  
John N. Klironomos ◽  
Miranda M. Hart
Keyword(s):  
Least Squares ◽  
Partial Least Squares ◽  
Ridge Regression ◽  
Repeated Measures ◽  
Partial Least Squares Regression ◽  
Plant Traits ◽  
Plant Biomass ◽  
Biomass Estimation ◽  
Least Squares Regression ◽  
Over Time

Estimating primary productivity over time is challenging for plant ecologists. The most accurate biomass measurements require destructive sampling and weighing. This is often not possible for manipulative studies that involve repeated measures over time, or for studies in protected areas. Estimates of aboveground plant biomass using allometric equations or linear regression on single plant traits have been used, but tend to have poor predictability both within and across systems, or are limited to specific plant taxa. Here we estimate aboveground plant biomass using multiple collinear plant traits to generate a standard curve specific to the site of interest. Partial least squares regression (PLS) and ridge regression (RR), addressing predictor collinearity, are robust, highly accurate statistical methods to estimate plant biomass across gross differences in plant morphology and growth habit.

scholarly journals Clinical characterization of dysautonomia in long COVID-19 patients

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Nicolas Barizien ◽  
Morgan Le Guen ◽  
Stéphanie Russel ◽  
Pauline Touche ◽  
Florent Huang ◽  
...  
Keyword(s):  
Mixed Models ◽  
Repeated Measures ◽  
Physical Medicine ◽  
Persistent Symptoms ◽  
Index Changes ◽  
Changes Over Time ◽  
And Control ◽  
Testing Group ◽  
Over Time

AbstractIncreasing numbers of COVID-19 patients, continue to experience symptoms months after recovering from mild cases of COVID-19. Amongst these symptoms, several are related to neurological manifestations, including fatigue, anosmia, hypogeusia, headaches and hypoxia. However, the involvement of the autonomic nervous system, expressed by a dysautonomia, which can aggregate all these neurological symptoms has not been prominently reported. Here, we hypothesize that dysautonomia, could occur in secondary COVID-19 infection, also referred to as “long COVID” infection. 39 participants were included from December 2020 to January 2021 for assessment by the Department of physical medicine to enhance their physical capabilities: 12 participants with COVID-19 diagnosis and fatigue, 15 participants with COVID-19 diagnosis without fatigue and 12 control participants without COVID-19 diagnosis and without fatigue. Heart rate variability (HRV) during a change in position is commonly measured to diagnose autonomic dysregulation. In this cohort, to reflect HRV, parasympathetic/sympathetic balance was estimated using the NOL index, a multiparameter artificial intelligence-driven index calculated from extracted physiological signals by the PMD-200 pain monitoring system. Repeated-measures mixed-models testing group effect were performed to analyze NOL index changes over time between groups. A significant NOL index dissociation over time between long COVID-19 participants with fatigue and control participants was observed (p = 0.046). A trend towards significant NOL index dissociation over time was observed between long COVID-19 participants without fatigue and control participants (p = 0.109). No difference over time was observed between the two groups of long COVID-19 participants (p = 0.904). Long COVID-19 participants with fatigue may exhibit a dysautonomia characterized by dysregulation of the HRV, that is reflected by the NOL index measurements, compared to control participants. Dysautonomia may explain the persistent symptoms observed in long COVID-19 patients, such as fatigue and hypoxia. Trial registration: The study was approved by the Foch IRB: IRB00012437 (Approval Number: 20-12-02) on December 16, 2020.

Sign in / Sign up

Export Citation Format

Share Document