run_cross_validation: Run k fold cross validation

In mdkessler/kTSCR: Implements k-Top Scoring Cluster Regression

Description

This function runs k fold cross validation by splitting input data into k partitions and holding out each partition as the test set in k different learning iterations.

Usage

run_cross_validation(
  y,
  X,
  Verbose.pass = FALSE,
  restrict = FALSE,
  rank = FALSE,
  Verbose = TRUE,
  standardize_features = TRUE,
  cluster_corr_prop = 1,
  ct = 1,
  sibling_prune = 0.1,
  k = 5,
  condensed_output = TRUE
)

Arguments

y	outcome variable
X	inut feature matrix
Verbose.pass	logical as to whether the kTSCR procedure should be verbose (i.e. should run_cross_validation pass 'verbose=TRUE' to get_top_clusters()) )
restrict	a list of colnames of X by which to restrict the analysis
rank	logical as to whether to use rank of outcome
Verbose	logical as to whether to be verbose
standardize_features	logical as to whether to standardize all features of X
cluster_corr_prop	what proportion of the maximum (weighted) cluster correlation with y should be reflected by the chosen siblings. A hyperparameter. Default is 1 (meaning include all elder-sibling pairs in cluster)
ct	correlation threshold determined how much a new cluster must improve the current correlation with y in order to be added as a top cluster. A hyperparameter. Default is 1 (meaning any improvement is sufficient to add the next cluster within the greedy framework)
sibling_prune	numeric between 0-1 that sets the threshold for how close apparent correlation and test correlation must be for a k-cv iteration to contribute its siblings to the final chosen siblings. In other words, a lower number is more stringent, since it means the overfitting had to be really low in a k-cv iteration for it to contribute to the final sibling output.
k	the k parameter in k fold cross validation (i.e. train/test partitions). Default is 5
condensed_output	return output that is condensed and summarized across k_cv iterations, specifically with regard to feature importance

Value

returns the list given by get_top_clusters for each n fold k cv run and includes the test correlation and train/test splits from each iteration

Examples

C <- 100  # represents samples
R <- 200 # represents features
y <- rnorm(C) # represents outcome variable
X <- matrix(rbeta(R*C, 2, 3), nrow = R)  # simulate data matrix
cv_res <- run_cross_validation

mdkessler/kTSCR documentation built on Feb. 25, 2021, 10:31 p.m.