ridge_regression: ridge_regression

In glucodensities/biosensors.usc: Distributional Data Analysis Techniques for Biosensor Data

ridge_regression

Description

Performs a Ridge regression.

Usage

ridge_regression(data, response, w=NULL, method="manhattan", type="gaussian")

Arguments

data	A biosensor object.
response	The name of the scalar response. The response must be a column name in data$variables.
w	A weight function.
method	The distance measure to be used (@seealso parallelDist::parDist). By default manhattan distance.
type	The kernel type ("gaussian" or "lapla"). By default gaussian distance.

Value

An object containing the components: best_alphas Best coefficients obtained with leave-one-out cross-validation criteria. best_kernel The kernel matrix of the best solution. best_sigma The sigma parameter of the best solution. best_lambda The lambda parameter of the best solution. sigmas The sigma parameters used in the fitting according to the median heuristic fitting criteria. predictions A matrix of predictions. r2 R-square of the different models fitted. error Mean squared-error of the different models fitted. predictions_cross A matrix of predictions obtained with leave-one-out cross-validation criteria.

Examples

# Data extracted from the paper: Hall, H., Perelman, D., Breschi, A., Limcaoco, P., Kellogg, R.,
# McLaughlin, T., Snyder, M., Glucotypes reveal new patterns of glucose dysregulation, PLoS
# biology 16(7), 2018.
file1 = system.file("extdata", "data_1.csv", package = "biosensors.usc")
file2 = system.file("extdata", "variables_1.csv", package = "biosensors.usc")
data = load_data(file1, file2)
regm = ridge_regression(data, "BMI")

glucodensities/biosensors.usc documentation built on May 12, 2022, 1:57 a.m.