singlesample_make_locto_splits.Rd

% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/singlesample-method-bank.R
\name{singlesample_make_locto_splits}
\alias{singlesample_make_locto_splits}
\title{Build leave-one-cancer-type-out splits with same-block exclusion}
\usage{
singlesample_make_locto_splits(
  sample_meta,
  cancer_col = "cancer_type",
  cohort_col = "cohort",
  provenance_block_col = "provenance_block",
  min_train_cohorts = 1L
)
}
\arguments{
\item{sample_meta}{Data frame with one row per sample.}

\item{cancer_col}{Column identifying cancer type or endpoint family.}

\item{cohort_col}{Column identifying cohorts.}

\item{provenance_block_col}{Optional column identifying specimen-provenance
blocks. When present, samples in the held-out block are excluded from
training.}

\item{min_train_cohorts}{Minimum number of training cohorts required for an
evaluable split.}
}
\value{
A list of split objects with train/test indices and status.
}
\description{
Build leave-one-cancer-type-out splits with same-block exclusion
}