Sieve-package {Sieve}

Nonparametric Estimation by the Method of Sieves

Description

Performs multivariate nonparametric regression/classification by the method of sieves (using orthogonal basis). The method is suitable for moderate high-dimensional features (dimension < 100). The l1-penalized sieve estimator, a nonparametric generalization of Lasso, is adaptive to the feature dimension with provable theoretical guarantees. We also include a nonparametric stochastic gradient descent estimator, Sieve-SGD, for online or large scale batch problems. Details of the methods can be found in: <arXiv:2206.02994> <arXiv:2104.00846><arXiv:2310.12140>.

Details

The DESCRIPTION file:

Package:	Sieve
Type:	Package
Title:	Nonparametric Estimation by the Method of Sieves
Version:	2.1
Date:	2023-10-19
Author:	Tianyu Zhang
Maintainer:	Tianyu Zhang <[e-mail]>
Description:	Performs multivariate nonparametric regression/classification by the method of sieves (using orthogonal basis). The method is suitable for moderate high-dimensional features (dimension < 100). The l1-penalized sieve estimator, a nonparametric generalization of Lasso, is adaptive to the feature dimension with provable theoretical guarantees. We also include a nonparametric stochastic gradient descent estimator, Sieve-SGD, for online or large scale batch problems. Details of the methods can be found in: <arXiv:2206.02994> <arXiv:2104.00846><arXiv:2310.12140>.
License:	GPL-2
Imports:	Rcpp, combinat, glmnet, methods, MASS
LinkingTo:	Rcpp, RcppArmadillo
RoxygenNote:	7.2.3
Encoding:	UTF-8

Index of help topics:

GenSamples              Generate some simulation/testing samples with
                        nonlinear truth.
Sieve-package           Nonparametric Estimation by the Method of
                        Sieves
clean_up_result         Clean up the fitted model
create_index_matrix     Create the index matrix for multivariate
                        regression
sieve.sgd.predict       Sieve-SGD makes prediction with new predictors.
sieve.sgd.preprocess    Preprocess the original data for sieve-SGD
                        estimation.
sieve.sgd.solver        Fit sieve-SGD estimators, using progressive
                        validation for hyperparameter tuning.
sieve_predict           Predict the outcome of interest for new samples
sieve_preprocess        Preprocess the original data for sieve
                        estimation.
sieve_solver            Calculate the coefficients for the basis
                        functions

~~ An overview of how to use the ~~ ~~ package, including the most ~~ ~~ important functions ~~

Author(s)

Tianyu Zhang

Maintainer: Tianyu Zhang <[e-mail]>

References

Tianyu Zhang and Noah Simon (2022) <arXiv:2206.02994>

Examples

xdim <- 5
basisN <- 1000
type <- 'cosine'

#non-linear additive truth. Half of the features are truly associated with the outcome
TrainData <- GenSamples(s.size = 300, xdim = xdim, 
            frho = 'additive', frho.para = xdim/2)

#noise-free testing samples
TestData <- GenSamples(s.size = 1e3, xdim = xdim, noise.para = 0, 
            frho = 'additive', frho.para = xdim/2)

sieve.model <- sieve_preprocess(X = TrainData[,2:(xdim+1)], 
            basisN = basisN, type = type, interaction_order = 2)

sieve.model <- sieve_solver(sieve.model, TrainData$Y, l1 = TRUE)

sieve_model_prediction <- sieve_predict(testX = TestData[,2:(xdim+1)], 
                testY = TestData$Y, sieve.model)