Scott Jennings (594642), David Lumpkin (11890475), Nils Warnock (11890478) et al. · Figshare · 2021
<p>Model selection results for evaluating differences in foraging habitat selection among wetland habitat types, while accounting for water depth, by GPS-tagged great egrets at Tomales Bay, CA, 2017–2020. K is the number of parameters, Δ AICc is the difference in AICc value between the top model and the current model, and AICc Wt. is the AICc model weight.</p> <p>(DOCX)</p>
Cited by 0Open access
Jean‐Marie Dufour · The New Palgrave Dictionary of Economics · 2018
The problem of statistical model selection in econometrics and statistics is reviewed. Model selection is interpreted as a decision problem through which a statistical model is selected in order to perform statistical analysis, such as estimation, testing, confidence set construction, forecasting, simulation, policy analysis, and so on. Broad approaches to model selection are described: (1) hypothesis testing procedures, including specification and diagnostic tests; (2) penalized goodness-of-fit methods, such as information criteria; (3) Bayesian approaches; (4) forecast evaluation methods. The effect of model selection on subsequent statistical inference is also discussed.
Cited by 0
Leonhard Held, Daniel Sabanés Bové · 2013
Cited by 1
Linhart, Heinz, Zucchini, Walter · 1986
Cited by 0
Abel Endawkie (13883778), Natnael Kebede (13142159), Desale Bihonegn Asmamaw (15338739) et al. · Figshare · 2024
<div><p>Background</p><p>Antenatal care (ANC) is an important component of maternal and child health care. To reduce perinatal mortality and improve women’s experience of care, the World Health Organization (WHO) recommends pregnant women should attend at least eight ANC visits. However, in Sub-Saharan Africa (SSA), the mean number of ANC visits among reproductive-age women using Demographic and Health Surveys (DHS) data following the new WHO recommendation is so far limited. Therefore, the study aimed to determine the mean number of ANC visits and its predictors among reproductive-age women in SSA.</p><p>Method</p><p>Community-based cross-sectional study was conducted among 188,880 weighted reproductive-age women in SSA using a recent round of DHS data from 2017–2023. Zero-inflated negative binomial regression (ZINB) was conducted and statistical significance was declared at p-value <0.05 and adjusted incidence rate ratio(AIRR) for count model part and adjusted odds ratio for logit model inflated part of ZINBR with 95% confidence interval (CI) were reported.</p><p>Result</p><p>The mean number of ANC visits among reproductive-age women in SSA was 4.08 with 95%CI [4.07, 4.09]. Among reproductive-age women who gave birth in the last five years before the survey, 7.3% had eight or more ANC visits during pregnancy. Age of women, maternal and husband educational status, types of pregnancy, birth order, household size, number of under-five children, and wealth index were associated with the numbers of ANC visits among reproductive-age women in Sub-Saharan.</p><p>Conclusion</p><p>The mean number of ANC visits among reproductive-age women in SSA is too lower than the new WHO recommendation of ANC visits for a positive pregnancy experience. This study also highlights that the proportion of at least eight ANC visits is low and there are still disparities in the mean of ANC visits across different regions of SSA. The increasing maternal age, higher maternal and husband educational status, wanted pregnancy, the number of household members, the number of under-five children, and higher wealth index increase the number of ANC visits. Unwanted pregnancy, no more fertility desire, and rural residences were contributed for zero ANC visits in SSA. Therefore, efforts should be geared towards improving maternal and husband’s educational status. We strongly recommend that the governments of SSA countries should empower women economically and educationally to achieve the goals of ANC as recommended by the WHO.</p></div>
Cited by 0Open access
Zeinah Elhaj Baddar (8905445), Erin Peck (10584182), Xiaoyu Xu (356293) · Figshare · 2021
Cited by 0Open access
Barbara Bravi (8520324), Vinod P. Balachandran (5565923), Benjamin D. Greenbaum (11379038) et al. · Figshare · 2021
<p>A: Hyperparametric search for training SONIA on count-weighted datasets: we set the <i>L</i><sub>2</sub> regularization at the intermediate value 0.01 (black dot), which ensures good performance at capturing both clone abundance (through ) and clonal fold change (through ) in the validation set. B. Data statistics (single-site frequency and connected correlations) are reproduced by SONIA. SONIA is documented to perform best when trained on lists of unique sequences [<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1009297#pcbi.1009297.ref015" target="_blank">15</a>], where also correlations are well reproduced (see inset). As a comparison, the gray points stand for the marginals given by the generation model <i>P</i><sub>gen</sub>. C-D: Hyperparametric search for training RBM on count-weighted datasets. Similarly to A, it is performed based on how clonal abundances and clone fold change are reproduced by and . We identify as optimal the combinations of parameters indicated by the black box, which result in maximal correlations with fold change (D) while keeping the correlation to abundance in the intermediate-high range (C). F-G: Hyperparametric search for training RBM on datasets of only responding CDR3s. Optimal <i>N</i><sup><i>h</i></sup> and (black dot) are chosen to ensure high likelihood on the training set and to prevent overfitting on the validation set. E-H show that in both cases single-site amino acid frequencies and connected correlations are extremely well reproduced by the RBM (with parameters marked by the black dot in D and G). In all hyperparametric searches shown, the model training set consist of 80% of all sequences in the dataset, the validation set is the remaining 20%. Datasets used: Pt3 NA from [<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1009297#pcbi.1009297.ref016" target="_blank">16</a>] (A-E), samples from Ref. [<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1009297#pcbi.1009297.ref017" target="_blank">17</a>] (F-H).</p> <p>(PNG)</p>
Cited by 0Open access
Sun Yue (321788), BingXiang Li (18980965) · Figshare · 2024
<div><p>Sustainable development and raising the domestic value-added rate of exports (DVARE) have become essential priorities in the pursuit of high-quality economic growth. An econometric spatial model is developed in this research using data on Chinese enterprises spanning 2008 to 2019. According to a study, exports’ domestic value-added rate (DVAR) can be successfully increased using environmental, social, and governance (ESG) practices. Spatial spillover analysis demonstrates that adopting ESG practices boosts export DVAR both within and between regions. According to heterogeneity analysis, the sample’s overall increase in DVARE as a result of ESG practices is mostly attributable to the mix and processing trade organizations, the eastern area, and large firms. An examination of the underlying mechanisms shows that businesses that implement advanced technologies are able to reinforce the favorable impact of ESG practices on DVARE. This article gives evidence from real-world studies that show how ESG practices help boost Chinese exports and advance sustainable development. The findings hold significant implications for other developing nations as they make the transition towards a pattern of economic growth.</p></div>
Cited by 0Open access
2004
Before using a parametri\n model one has to be sure that it oers a reason-able des\nription of the system to be modeled. If a bad model stru\nture is employed, the obtained model will also be bad, no matter how good is the parameter estimation method. There exist many possible ways of validating andidate models. This thesis fo\nuses on one of the most \nommon ways, i.e., the use of information \nriteria. First, some \nommon information \nriteria are presented, and in the later \nhapters, various extentions and implementa-tions are shown. An important extention, whi\nh is advo\nated in the thesis, is the multi-model (or model averaging) approa\nh to model sele\ntion. This multi-model approa\nh \nonsists of forming a weighted sum of several \nandi-date models, whi\nh then \nan be used for inferen\ne.
Cited by 0
Hannes Leeb, Benedikt M. Pötscher · 2008
Cited by 0
Jean‐Marie Dufour · The New Palgrave Dictionary of Economics · 2008
The problem of statistical model selection in econometrics and statistics is reviewed. Model selection is interpreted as a decision problem through which a statistical model is selected in order to perform statistical analysis, such as estimation, testing, confidence set construction, forecasting, simulation, policy analysis, and so on. Broad approaches to model selection are described: (1) hypothesis testing procedures, including specification and diagnostic tests; (2) penalized goodness-of-fit methods, such as information criteria; (3) Bayesian approaches; (4) forecast evaluation methods. The effect of model selection on subsequent statistical inference is also discussed.
Cited by 0
The MIT Press eBooks · 2012
Cited by 0
Rolf Wuthrich, Carole El Ayoubi · 2025
Cited by 0
Selén, Yngve · KTH Publication Database DiVA (KTH Royal Institute of Technology) · 2004
Before using a parametric model one has to be sure that it offers a reasonable description of the system to be modeled. If a bad model structure is employed, the obtained model will also be bad, no matter how good is the parameter estimation method. There exist many possible ways of validating candidate models. This thesis focuses on one of the most common ways, i.e., the use of information criteria. First, some common information criteria are presented, and in the later chapters, various extentions and implementations are shown. An important extention, which is advocated in the thesis, is the multi-model (or model averaging) approach to model selection. This multi-model approach consists of forming a weighted sum of several candidate models, which then can be used for inference.
Cited by 0Open access
Martin Sewell · 2007
An introduction to model selection. Science is the systematic study of the universe—through observation and experiment—in the pursuit of knowledge that allows us to generalize. Although considered bad form in the current climate of political correctness (Lind 1998,
Cited by 0
Wolfgang von der Linden, V. Dose, U. von Toussaint · Cambridge University Press eBooks · 2014
A summary is not available for this content so a preview has been provided. Please use the Get access link above for information on how to access this content.
Cited by 0
WORLD SCIENTIFIC eBooks · 2021
Cited by 0
Heinz Linhart, Walter 1946- Zucchini · 1986
Cited by 0
Albert Zejda · Digital Repository (National Repository of Grey Literature) · 2025
This thesis deals with the specification of a class of adequate models, selected from the set of all considered models, for a given dataset. The aim is to present a procedure that minimizes the number of comparisons/tests needed to identify this subset of models, along with an analysis of the properties of the proposed procedure. The approach is based on the so-called principle of coherence. It utilizes a suitable algebraic structure in the form of a lattice over the space of all considered models. The result is an algorithm that, with respect to a specified decision rule, enables the identification of an appropriate class of models for the given data.
Cited by 0
Emily S. Nightingale (9148620), Lloyd A. C. Chapman (9106122), Sridhar Srikantiah (3357185) et al. · Figshare · 2020
Cited by 0Open access