المؤلفون: Xiaowei Zhang, Junliang Li

المصدر: AIMS Mathematics, Vol 9, Iss 6, Pp 16392-16421 (2024)

مصطلحات موضوعية: model averaging, causal inference, partially linear models, conditional average treatment effect, jackknife-type criterion, Mathematics, QA1-939

الوصف: Treatment effects with heterogeneity and heteroskedasticity are widely studied and applied in many fields, such as statistics and econometrics. The conditional average treatment effect provides an excellent measure of the heterogeneous treatment effect. In this paper, we propose a model averaging estimation for the conditional average treatment effect with partially linear models based on the jackknife-type criterion under heteroscedastic error. Within this context, we provide theoretical justification for our model averaging approach, and we establish asymptotic optimality and weight convergence properties for our model under certain conditions. The performance of our proposed estimator is compared with that of classical estimators by using a Monte Carlo study and empirical analysis.

وصف الملف: electronic resource

العلاقة: https://doaj.org/toc/2473-6988Test

الوصول الحر: https://doaj.org/article/ebc4f476be5a4d2f88899b3979d107dcTest

المؤلفون: Haiyan Su, Linlin Chen

المصدر: Mathematics, Vol 12, Iss 1, p 162 (2024)

مصطلحات موضوعية: partially linear models, empirical likelihood, confidence interval, Mathematics, QA1-939

الوصف: Partially linear models find extensive application in biometrics, econometrics, social sciences, and various other fields due to their versatility in accommodating both parametric and nonparametric elements. This study aims to establish statistical inference for the parametric component effects within these models, employing a nonparametric empirical likelihood approach. The proposed method involves a projection step to eliminate the nuisance nonparametric component and utilizes an empirical-likelihood-based technique, along with the Bartlett correction, to enhance the coverage probability of the confidence interval for the parameter of interest. This method demonstrates robustness in handling normally and non-normally distributed errors. The proposed empirical likelihood ratio statistic converges to a limiting chi-square distribution under certain regulations. Simulation studies demonstrate that this method provides better inference in terms of coverage probabilities compared to the conventional normal-approximation-based method. The proposed method is illustrated by analyzing the Boston housing data from a real study.

العلاقة: https://www.mdpi.com/2227-7390/12/1/162Test; https://doaj.org/toc/2227-7390Test; https://doaj.org/article/5795f1bfb8f94de888c23da294f4af7cTest

الإتاحة: https://doi.org/10.3390/math12010162Test
https://doaj.org/article/5795f1bfb8f94de888c23da294f4af7cTest

المؤلفون: Yanting Xiao, Wanying Dong

المصدر: AIMS Mathematics, Vol 8, Iss 8, Pp 18373-18391 (2023)

مصطلحات موضوعية: varying-coefficient partially linear models, auxiliary variables, error-in-variable, modal regression, Mathematics, QA1-939

الوصف: We study the varying-coefficient partially linear model when some linear covariates are not observed, but their auxiliary instrumental variables are available. Combining the calibrated error-prone covariates and modal regression, we present a two-stage efficient estimation procedure, which is robust against outliers or heavy-tail error distributions. Asymptotic properties of the resulting estimators are established. Performance of our proposed estimation procedure is illustrated through some numerous simulations and a real example. And the results confirm that the proposed methods are satisfactory.

وصف الملف: electronic resource

العلاقة: https://doaj.org/toc/2473-6988Test

الوصول الحر: https://doaj.org/article/3b5188b418ea407fb635bca25c248b95Test

المؤلفون: Xin Ma, Yubin Cai, Hong Yuan, Yanqiao Deng

المصدر: Sustainability; Volume 15; Issue 9; Pages: 7086

مصطلحات موضوعية: support vector machines, principal component analysis, partially linear models, primary energy consumption

جغرافية الموضوع: agris

الوصف: Energy forecasting based on univariate time series has long been a challenge in energy engineering and has become one of the most popular tasks in data analytics. In order to take advantage of the characteristics of observed data, a partially linear model is proposed based on principal component analysis and support vector machine methods. The principal linear components of the input with lower dimensions are used as the linear part, while the nonlinear part is expressed by the kernel function. The primal-dual method is used to construct the convex optimization problem for the proposed model, and the sequential minimization optimization algorithm is used to train the model with global convergence. The univariate forecasting scheme is designed to forecast the primary energy consumption of the electric power sector of the United States using real-world data sets ranging from January 1973 to January 2020, and the model is compared with eight commonly used machine learning models as well as the linear auto-regressive model. Comprehensive comparisons with multiple evaluation criteria (including 19 metrics) show that the proposed model outperforms all other models in all scenarios of mid-/long-term forecasting, indicating its high potential in primary energy consumption forecasting.

وصف الملف: application/pdf

العلاقة: Energy Sustainability; https://dx.doi.org/10.3390/su15097086Test

الإتاحة: https://doi.org/10.3390/su15097086Test

المؤلفون: Bal, Çağatay, Yılmaz, Ersin

المساهمون: MÜ, Fen Fakültesi, İstatistik Bölümü, orcid:0000-0002-7823-2712, orcid:0000-0002-9871-4700, Bal, Çağatay, Yılmaz, Ersin

مصطلحات موضوعية: Feed forward neural networks, Smoothing splines, Partially linear models, Right-censored data, İleri beslemeli sinir ağları, Splayn düzleştirme, Kısmi doğrusal modeller, Sağdan-sansürlü veri

الوصف: Objective: Modeling right-censored data becomes a challenging task in survival analysis, due to having an incomplete data structure. When the response variable is rightcensored, classical estimation methods cannot be used directly. Therefore, the censorship problem should be solved before the modeling process. The purpose of this study is to solve the censorship problem with synthetic data transformation and to make a comparison between the partial linear model (PLM) and feed forward neural networks (FFNN), two popular modeling procedures in recent years, in terms of model residuals. Thus, it is to study the behavior of methods. Material and Methods: This paper aims to estimate the effects of explanatory variables on a right-censored response variable whose distribution is unknown by two different methods, PLM and FFNN based on the spline smoothing method. The spline smoothing method is a mathematical approximation method used in PLM estimation. FFNN is a machine learning method that has become very popular recently and produces satisfactory models. To overcome the censorship problem, the right-censored response variable for the two mentioned methods has been replaced with synthetic data. Synthetic data transformation is a widely used censorship resolution method that allows censorship presence to be included in the prediction process. Results: To achieve the aim of the study, both simulation and real data studies were carried out and the results were presented. Conclusion: Kidney weakness data is used as an example of real data. When the results are examined, it is seen that FFNN is superior to PLM in both numerical studies. ; Amaç: Eksik bir veri yapısına sahip olması nedeniyle sağdansansürlenmiş verilerin modellenmesi, sağkalım analizinde zor bir işlemdir. Yanıt değişkeni sağdan sansürlendiğinde, klasik tahmin yöntemleri doğrudan kullanılamaz. Bu nedenle modelleme sürecinden önce sansür problemi çözülmelidir. Bu çalışmanın amacı, sansür problemini sentetik veri dönüşüm ile çözerek literatürde son yıllarda ...

وصف الملف: application/pdf

العلاقة: Türkiye Klinikleri Biyoistatistik Dergisi; Makale - Uluslararası Hakemli Dergi - Kurum Öğretim Elemanı; BAL C, YILMAZ E (2022). Modelling Right-Censored Data with Partially Linear Model and Feed Forward Neural Networks: A Methodological Study. Türkiye Klinikleri Biyoistatistik Dergisi, 14(2), 90 - 102. 10.5336/biostatic.2022-89354; 1308-7894 / 2146-8877; file:///C:/Users/Aidata/Downloads/document-39.pdf; https://hdl.handle.net/20.500.12809/10526Test; 14; 90; 102

الإتاحة: https://doi.org/20.500.12809/10526Test
https://hdl.handle.net/20.500.12809/10526Test

المؤلفون: Simó, Amelia, Ibáñez Gual, Maria Victoria, Epifanio, Irene, Gimeno, Vicent

مصطلحات موضوعية: children's wear, generalized linear models, Kernel regression, partially linear models, shape space, statistical shape analysis

الوصف: We introduce generalized partially linear models with covariates on Riemannian manifolds. These models, like ordinary generalized linear models, are a generalization of partially linear models on Riemannian manifolds that allow for scalar response variables with error distribution models other than a normal distribution. Partially linear models are particularly useful when some of the covariates of the model are elements of a Riemannian manifold, because the curvature of these spaces makes it difficult to define parametric models. The model was developed to address an interesting application: the prediction of children's garment fit based on three‐dimensional scanning of their bodies. For this reason, we focus on logistic and ordinal models and on the important and difficult case where the Riemannian manifold is the three‐dimensional case of Kendall's shape space. An experimental study with a well‐known three‐dimensional database is carried out to check the goodness of the procedure. Finally, it is applied to a three‐dimensional database obtained from an anthropometric survey of the Spanish child population. A comparative study with related techniques is carried out.

وصف الملف: application/pdf

العلاقة: Spanish Ministry of Economy and Competitiveness with Federación Española de Enfermedades Raras fundsand (grant DPI2017-87333-R); Universitat Jaume I (grant UJI-B2017-13); SIMÓ, Amelia, et al. Generalized partially linear models on Riemannian manifolds. JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES C-APPLIED STATISTICS 2020, 69(3): 641-661; http://hdl.handle.net/10234/188460Test; https://doi.org/10.1111/rssc.12411Test

الإتاحة: https://doi.org/10.1111/rssc.12411Test
http://hdl.handle.net/10234/188460Test

المؤلفون: Bal, Çağatay, Yılmaz, Ersin

المساهمون: MÜ, Fen Fakültesi, İstatistik Bölümü, Bal, Çağatay, Yılmaz, Ersin

المصدر: Turkiye Klinikleri Journal of Biostatistics. 14:90-102

مصطلحات موضوعية: Smoothing splines, Feed forward neural networks, Splayn düzleştirme, İleri beslemeli sinir ağları, Right-censored data, General Medicine, Kısmi doğrusal modeller, Partially linear models, Sağdan-sansürlü veri

الوصف: Objective: Modeling right-censored data becomes a challenging task in survival analysis, due to having an incomplete data structure. When the response variable is rightcensored, classical estimation methods cannot be used directly. Therefore, the censorship problem should be solved before the modeling process. The purpose of this study is to solve the censorship problem with synthetic data transformation and to make a comparison between the partial linear model (PLM) and feed forward neural networks (FFNN), two popular modeling procedures in recent years, in terms of model residuals. Thus, it is to study the behavior of methods. Material and Methods: This paper aims to estimate the effects of explanatory variables on a right-censored response variable whose distribution is unknown by two different methods, PLM and FFNN based on the spline smoothing method. The spline smoothing method is a mathematical approximation method used in PLM estimation. FFNN is a machine learning method that has become very popular recently and produces satisfactory models. To overcome the censorship problem, the right-censored response variable for the two mentioned methods has been replaced with synthetic data. Synthetic data transformation is a widely used censorship resolution method that allows censorship presence to be included in the prediction process. Results: To achieve the aim of the study, both simulation and real data studies were carried out and the results were presented. Conclusion: Kidney weakness data is used as an example of real data. When the results are examined, it is seen that FFNN is superior to PLM in both numerical studies. Amaç: Eksik bir veri yapısına sahip olması nedeniyle sağdansansürlenmiş verilerin modellenmesi, sağkalım analizinde zor bir işlemdir. Yanıt değişkeni sağdan sansürlendiğinde, klasik tahmin yöntemleri doğrudan kullanılamaz. Bu nedenle modelleme sürecinden önce sansür problemi çözülmelidir. Bu çalışmanın amacı, sansür problemini sentetik veri dönüşüm ile çözerek literatürde son yıllarda popüler olarak kullanılan 2 modelleme prosedürü olan kısmi doğrusal model [partial linear model (PLM)] ve ileri beslemeli sinir ağları [feed forward neural networks (FFNN)] arasında model artıkları açısından bir karşılaştırma yapmak ve böylece yöntemlerin davranışlarını incelemektir. Gereç ve Yöntemler: Bu makale, açıklayıcı değişkenlerin, dağılımı bilinmeyen sağdan sansürlü bir yanıt değişkeni üzerindeki etkilerini, splayn düzleştirme yöntemine dayalı PLM ve FFNN olmak üzere 2 farklı yöntemle tahmin etmeyi amaçlar. Splayn düzleştirme yöntemi, PLM tahmininde kullanılan matematiksel yaklaştırma yöntemidir. FFNN ise son zamanlarda oldukça popülerleşen ve tatmin edici modeller üreten bir makine öğrenmesi yöntemidir. Sansür sorununun üstesinden gelmek için bahsedilen 2 yöntem için sağdan sansürlü yanıt değişkeni sentetik verilerle değiştirilmiştir. Sentetik veri dönüşümü, sansür varlığını tahmin sürecine dâhil edilmesini sağlayan yaygın kullanılan bir sansür çözüm yöntemidir. Bulgular: Çalışmanın amacına ulaşmak için hem simülasyon hem de gerçek veri çalışmaları yapılmış ve sonuçlar sunulmuştur. Sonuç: Gerçek veri örneği olarak böbrek zayıflığı verisi kullanılmıştır. Sonuçlar incelendiğinde, her iki sayısal çalışmada da FFNN’nin PLM’ye üstünlük sağladığı açıkça görülmektedir.

وصف الملف: application/pdf

الوصول الحر: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::575bd7447c28ce4136de40bc85dfc5d8Test
https://doi.org/10.5336/biostatic.2022-89354Test

المؤلفون: Fan, Jianqing, Huang, Tao

المصدر: Bernoulli, 2005 Dec 01. 11(6), 1031-1057.

الوصول الحر: https://www.jstor.org/stable/25464778Test

المؤلفون: Sato, Takaki

المصدر: DSSR Discussion Papers. 104:1-23

مصطلحات موضوعية: Series estimation, Instrumental variables, Partially linear models, Spatial econometrics

الوصول الحر: https://explore.openaire.eu/search/publication?articleId=jairo_______::c56571ede0344e76fef1c88a1ce90aa3Test
http://hdl.handle.net/10097/00126434Test

المؤلفون: Nuñez Lemus, Marcela, 1989

المساهمون: Matos, Larissa Avila, 1987, Lachos Dávila, Víctor Hugo, 1973, Ludwig, Guilherme Vieira Nunes, Ferreira, Clecio da Silva, Universidade Estadual de Campinas. Instituto de Matemática, Estatística e Computação Científica, Programa de Pós-Graduação em Estatística, UNIVERSIDADE ESTADUAL DE CAMPINAS

المصدر: Biblioteca Digital de Teses e Dissertações da Universidade Estadual de Campinas (UNICAMP)
Universidade Estadual de Campinas (UNICAMP)
instacron:UNICAMP

مصطلحات موضوعية: Local influence (Statistics), Scale mixtures, Censored observations (Statistics), Influência local (Estatística), Modelos lineares parciais, Algoritmos de esperança-maximização, Misturas de escala (Estatística), Expectation-maximization algorithms, Partially linear models, Observações censuradas (Estatística)

الوصف: Orientadores: Larissa Avila Matos, Víctor Hugo Lachos Dávila Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Matemática Estatística e Computação Científica Resumo: Em muitos estudos, dados limitados ou censurados são coletados. Isso ocorre em várias situações práticas, devido as limitações dos equipamentos de medição ou pelo desenho experimental. Dessa forma, as respostas podem ser censuradas à esquerda, à direita ou em um intervalo. Por outro lado, os modelos parcialmente lineares são considerados como uma extensão flexível dos modelos de regressão lineares incluindo uma componente não paramétrica em alguma covariável. Neste trabalho, estudamos procedimentos de estimação e diagnóstico em modelos de regressão parcialmente lineares com respostas censuradas sob a classe de distribuições de mistura de escala normal (SMN). Esta família de distribuições contém um grupo de distribuições com caudas mais pesadas do que a normal que costumam ser usadas para inferências robustas de dados simétricos, como a t de Student, a slash, a normal contaminada, entre outras. Um algoritmo do tipo EM é apresentado para obter iterativamente as estimativas de máxima verossimilhança penalizada dos parâmetros dos modelos. Para examinar o desempenho dos modelos propostos, técnicas de deleção de casos e de influência local são desenvolvidas para mostrar a robustez contra observações potencialmente influentes e outliers. Isto é feito através da análise de sensibilidade das estimativas de máxima verossimilhança penalizada com alguns esquemas de perturbação no modelo ou nos dados e analisando alguns gráficos de diagnóstico. A eficácia do método proposto é avaliada através da análise de conjuntos de dados simulados e reais. O pacote \verb+PartCensReg+ implementado no R dá suporte computacional para este trabalho Abstract: In many studies, limited or censored data are collected. This occurs, in many situations in practice, for reasons such as limitations of measuring instruments or due to experimental design. So, the responses can be either left, interval or right censored. On the other hand, partially linear models are considered as a flexible generalizations of linear regression models by including a nonparametric component of some covariate in the linear predictor. In this work, we discuss estimation and diagnostic procedures in partially linear censored regression models with errors following a scale mixture of normal (SMN) distributions. This family of distributions contains a group of well-known heavy-tailed distributions that are often used for robust inference of symmetrical data, such as Student-t, slash and contaminated normal, among others. A simple EM-type algorithm for iteratively computing maximum penalized likelihood (MPL) estimates of the parameters is presented. To examine the performance of the proposed model, case-deletion and local influence techniques are developed to show its robustness against outlying and influential observations. This is performed by sensitivity analysis of the maximum penalized likelihood estimates under some usual perturbation schemes, either in the model or in the data, and by inspecting some proposed diagnostic graphs. We evaluate the finite sample performance of the algorithm and the asymptotic properties of the MPL estimates through empirical experiments. An application to a real dataset is presented to illustrate the effectiveness of the proposed methods. The package \verb+PartCensReg+ implemented for the software R give computational support to this work Mestrado Estatística Mestre em Estatística CAPES

وصف الملف: application/pdf; 1 recurso online (70 p.) : il., digital, arquivo PDF.

الوصول الحر: https://explore.openaire.eu/search/publication?articleId=doi_dedup___::096890a8a020a620b795629e08407f1fTest
https://doi.org/10.47749/t/unicamp.2018.995802Test