Variable selection in partially linear additive hazards model with grouped covariates and a diverging number of parameters

Afzal, Arfan Raheen; Yang, Jing; Lu, Xuewen

doi:10.1007/s00180-020-01062-3

Cited by 3 publications

(3 citation statements)

References 63 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To conclude the asymptotic normality of the proposed estimators, we impose the nonlinear regression function φ is a B-spline function, where the knots and order are assumed to be known. The primary purpose is to avoid the complex derivative of the limiting property, and the same skill is also adopted in other areas, such as Afzal et al (2017Afzal et al ( ,2021. Even if this requirement does not hold strictly, the following simulation studies demonstrate that the approximation error can be ignorable.…”

Section: Estimation Methodsmentioning

confidence: 99%

“…Explicitly speaking, based on the partially linear proportional hazards model, Liu et al (2016) proposed a new penalised pseudo-partial likelihood method to select important covariates for multivariate failure time data. Afzal et al (2017) considered partly linear AH model for left-truncated and right-censored data, and recently Afzal et al (2021) proposed a hierarchical bi-level variable selection approach for right censored data in the linear part of this model, where the covariates are naturally grouped. Song et al (2019) considered a partially time-varying coefficient proportional hazards model, where corrected score and conditional score approaches are employed to accommodate potential measurement error.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Partially Linear Additive Hazards Regression for Clustered and Right Censored Data

Chen¹,

Ren²

2022

Bulletin of Informatics and Cybernetics

View full text Add to dashboard Cite

For analyzing clustered survival data, a flexible partially linear additive hazards model is proposed. To accommodate the nonlinear effects, the unknown regression function is approximated by B-splines. All regression coefficients are estimated through a system of pseudo-score functions. Under certain conditions, the proposed estimators are shown to be asymptotically normal, where a consistent estimator of the covariance matrix is given. Simulation studies are also conducted to evaluate the finite sample performance of the proposed method, which is illustrated using a real data set from an AIDS clinical trial.

show abstract

Section: Estimation Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Partially Linear Additive Hazards Regression for Clustered and Right Censored Data

Chen¹,

Ren²

2022

Bulletin of Informatics and Cybernetics

View full text Add to dashboard Cite

show abstract

“…Simpler techniques like the classical least absolute shrinkage and selection operator (LASSO) (Tibshirani, 1996) can mimic this to some extent if it is assumed that a group is selected as soon as at least one of its members is included in the model. It is still an unverified hypothesis that dedicated bi‐level selection procedures are superior to such a strategy in terms of interpretability (Gregorutti et al., 2015; Lee et al., 2018; Afzal et al., 2021). Since they are not designed to optimize collinearity tolerance or group‐level consistency, it remains unclear which characteristic should account for their supposed superiority.…”

Section: Introductionmentioning

confidence: 99%

Interpretability of bi‐level variable selection methods

Buch,

Schulz,

Schmidtmann

et al. 2024

Biometrical J

View full text Add to dashboard Cite

Variable selection is usually performed to increase interpretability, as sparser models are easier to understand than full models. However, a focus on sparsity is not always suitable, for example, when features are related due to contextual similarities or high correlations. Here, it may be more appropriate to identify groups and their predictive members, a task that can be accomplished with bi‐level selection procedures. To investigate whether such techniques lead to increased interpretability, group exponential LASSO (GEL), sparse group LASSO (SGL), composite minimax concave penalty (cMCP), and least absolute shrinkage, and selection operator (LASSO) as reference methods were used to select predictors in time‐to‐event, regression, and classification tasks in bootstrap samples from a cohort of 1001 patients. Different groupings based on prior knowledge, correlation structure, and random assignment were compared in terms of selection relevance, group consistency, and collinearity tolerance. The results show that bi‐level selection methods are superior to LASSO in all criteria. The cMCP demonstrated superiority in selection relevance, while SGL was convincing in group consistency. An all‐round capacity was achieved by GEL: the approach jointly selected correlated and content‐related predictors while maintaining high selection relevance. This method seems recommendable when variables are grouped, and interpretation is of primary interest.

show abstract

Group sparse sufficient dimension reduction: a model-free group variable selection method

Cai,

Lu,

Shen

2024

Comput Stat

View full text Add to dashboard Cite

Variable selection in partially linear additive hazards model with grouped covariates and a diverging number of parameters

Cited by 3 publications

References 63 publications

Partially Linear Additive Hazards Regression for Clustered and Right Censored Data

Partially Linear Additive Hazards Regression for Clustered and Right Censored Data

Interpretability of bi‐level variable selection methods

Group sparse sufficient dimension reduction: a model-free group variable selection method

Contact Info

Product

Resources

About