lincoef: Linear Model Coefficients

In egenn/rtemis: Machine Learning and Visualization

Linear Model Coefficients

Description

Get linear model coefficients

Usage

lincoef(
  x,
  y,
  weights = NULL,
  method = "glmnet",
  type = c("Regression", "Classification", "Survival"),
  learning.rate = 1,
  alpha = 1,
  lambda = 0.05,
  lambda.seq = NULL,
  cv.glmnet.nfolds = 5,
  which.cv.glmnet.lambda = c("lambda.min", "lambda.1se"),
  nbest = 1,
  nvmax = 8,
  sgd.model = "glm",
  sgd.model.control = list(lambda1 = 0, lambda2 = 0),
  sgd.control = list(method = "ai-sgd"),
  trace = 0
)

Arguments

x	Feature matrix or data.frame. Will be coerced to data.frame for method = "allSubsets", "forwardStepwise", or "backwardStepwise"
y	Outcome
weights	Float, vector: Case weights
method	Character: Method to use: "glm": uses stats::lm.wfit "glmnet": uses glmnet::glmnet "cv.glmnet": uses glmnet:cv.glmnet "lm.ridge": uses MASS::lm.ridge "allsubsets": uses leaps::regsubsets with method = "exhaustive" "forwardStepwise": uses leaps::regsubsets with method = "forward" "backwardStepwise": uses leaps::regsubsets with method = "backward" "sgd": uses sgd::sgd "solve": uses base::solve "none": fits no model and returns all zeroes, for programming convenience in special cases
type	Character: "Regression", "Classification", or "Survival"
learning.rate	Numeric: Coefficients will be multiplied by this number
alpha	Float: alpha for method = glmnet or cv.glmnet.
lambda	Float: The lambda value for glmnet, cv.glmnet, lm.ridge Note: For glmnet and cv.glmnet, this is the lambda used for prediction. Training uses lambda.seq.
lambda.seq	Float, vector: lambda sequence for glmnet and cv.glmnet.
cv.glmnet.nfolds	Integer: Number of folds for cv.glmnet
which.cv.glmnet.lambda	Character: Whitch lambda to pick from cv.glmnet: "lambda.min": Lambda that gives minimum cross-validated error;
nbest	Integer: For method = "allSubsets", number of subsets of each size to record. Default = 1
nvmax	Integer: For method = "allSubsets", maximum number of subsets to examine.
sgd.model	Character: Model to use for method = "sgd".
sgd.model.control	List: model.control list to pass to sgd::sgd
sgd.control	List: sgd.control list to pass to sgd::sgd "lambda.1se": Largest lambda such that error is within 1 s.e. of the minimum.
trace	Integer: If set to zero, all warnings are ignored

Details

This function minimizes checks for speed. It doesn't check dimensionality of x. Only use methods "glm", "sgd", or "solve" if there is only one feature in x.

Value

Named numeric vector of linear coefficients

Author(s)

E.D. Gennatas

egenn/rtemis documentation built on Oct. 28, 2024, 6:30 a.m.