Fit the combiner on the out-of-fold predictions

The combiner sees the out-of-fold predictions, the response, the mask of scorable cells and the fold map, and never a model. That is what keeps it honest: there is no way for it to read anything a candidate fitted in-sample, because it is not handed one.

Usage

ensemble_fit(oof, y, cells, folds, spec = ensemble(), scores = NULL)

Arguments

oof

Named list of [target, response] matrices, one per candidate.

y

The response matrix.

cells

The scorable-cell mask from scorable_cells().

folds

The fold map.

spec

An ensemble() specification.

scores

The per-cell scores of the run, a data frame carrying candidate, variable, fold, score and scorable. Read only where spec names no metric of its own.

Value

A timesift_stack: method, the named weights, and what they were fitted on.

Examples

set.seed(1)
y <- matrix(rbinom(200, 1, 0.4), nrow = 50,
            dimnames = list(sprintf("p%02d", 1:50), paste0("sp", 1:4)))
folds <- fold_map(y, v = 5)
truth <- matrix(runif(200), nrow = 50, dimnames = dimnames(y))
oof <- list(good = 0.8 * y + 0.2 * truth, noise = truth)
ensemble_fit(oof, y, scorable_cells(y, folds), folds)