Abstract: An envelope is a targeted dimension reduction subspace for simultaneously achieving dimension reduction and improving parameter estimation efficiency. While many envelope methods have been proposed in recent years, all envelope methods hinge on the knowledge of a key hyperparameter, the structural dimension of the envelope. How to estimate the envelope dimension consistently is of substantial interest from both theoretical and practical aspects. Moreover, very recent advances in the literature have generalized envelope as a model-free method, which makes selecting the envelope dimension even more challenging. Likelihood-based approaches such as information criteria and likelihood-ratio tests either cannot be directly applied or have no theoretical justification. To address this critical issue of dimension selection, we propose two unified approaches -- called FG and 1D selections -- for determining the envelope dimension that can be applied to any envelope models and methods. The two model-free selection approaches are based on the two different envelope optimization procedures: the full Grassmannian (FG) optimization and the 1D algorithm (Cook and Zhang, 2016), and are shown to be capable of correctly identifying the structural dimension with a probability tending to 1 under mild moment conditions as the sample size increases. While the FG selection unifies and generalizes the BIC and modified BIC approaches that existing in the literature, and hence provides the theoretical justification of them under weak moment condition and model-free context, the 1D selection is computationally more stable and efficient in finite sample. Extensive simulations and a real data analysis demonstrate the superb performance of our proposals.
Recommendations
- A theoretical view of the envelope model for multivariate linear regression as response dimension reduction
- Functional envelope for model-free sufficient dimension reduction
- Foundations for envelope models and methods
- Envelopes and reduced-rank regression
- An introduction to envelopes. Dimension reduction for efficient estimation in multivariate statistics
Cites work
- scientific article; zbMATH DE number 47995 (Why is no real title available?)
- A new look at the statistical model identification
- A useful variant of the Davis-Kahan theorem for statisticians
- A validated information criterion to determine the structural dimension in dimension reduction models
- Aster models for life history analysis
- Determining the Dimensionality in Sliced Inverse Regression
- Determining the dimension of iterative Hessian transformation
- Determining the dimension of the central subspace and central mean subspace
- Dimension reduction in regressions through cumulative slicing estimation
- Envelope models for parsimonious and efficient multivariate linear regression
- Envelopes and Partial Least Squares Regression
- Envelopes and reduced-rank regression
- Estimating the dimension of a model
- Fast envelope algorithms
- Foundations for envelope models and methods
- On Sliced Inverse Regression With High-Dimensional Covariates
- On the consistency of coordinate-independent sparse estimation with BIC
- On the likelihood ratio test for envelope models in multivariate linear regression
- Partial envelopes for efficient estimation in multivariate linear regression
- Rank estimation in reduced-rank regression
- Scaled envelopes: scale-invariant and efficient estimation in multivariate linear regression
- Weighted envelope estimation to handle variability in model selection
Cited in
(13)- Model-based clustering with envelopes
- Parsimonious Tensor Discriminant Analysis
- A note on hypothesis testing in the envelope model
- Frequentist model averaging for envelope models
- Envelope methods
- Efficient Integration of Sufficient Dimension Reduction and Prediction in Discriminant Analysis
- Tensor envelope mixture model for simultaneous clustering and multiway dimension reduction
- Envelopes and principal component regression
- A Review of Envelope Models
- New parsimonious multivariate spatial model: spatial envelope
- Combining envelope methodology and aster models for variance reduction in life history analyses
- General model-free weighted envelope estimation
- Enveloped Huber Regression
This page was built for publication: Model-free envelope dimension selection
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1657950)
