Skip to contents

Combine Multiple Channels

Source: R/step_channel_combine.R
step_measure_channel_combine.Rd

step_measure_channel_combine() creates a specification of a recipe step that combines multiple measurement channels into a single representation.

Usage

step_measure_channel_combine(
  recipe,
  ...,
  strategy = c("stack", "concat", "weighted_sum", "mean"),
  weights = NULL,
  output_col = ".measures",
  remove_original = TRUE,
  role = NA,
  trained = FALSE,
  skip = FALSE,
  id = recipes::rand_id("measure_channel_combine")
)

Arguments

recipe

A recipe object.

...

One or more selector functions to choose measure columns. If empty, all measure columns are used.

strategy

How to combine channels:

  • "stack": Stack channels into an nD measurement with channel as a dimension

  • "concat": Concatenate channels into a single 1D measurement

  • "weighted_sum": Compute weighted sum across channels

  • "mean": Average across channels (equal weights)

weights

For strategy = "weighted_sum", a numeric vector of weights. Must have same length as number of channels. Default is equal weights.

output_col

Name of the output measure column. Default is ".measures".

remove_original

Logical. Should original channel columns be removed? Default is TRUE.

role

Not used.

trained

Logical indicating if the step has been trained.

skip

Logical. Should the step be skipped when baking?

id

Unique step identifier.

Value

An updated recipe with the new step added.

Details

After aligning multiple channels to a common grid with step_measure_channel_align(), this step combines them for downstream analysis. The choice of strategy depends on the analysis goal:

Strategies

  • stack: Creates an n-dimensional measurement where channel becomes a dimension. Useful for multi-way analysis (PARAFAC, Tucker).

  • concat: Concatenates all channels end-to-end into a single long vector. Useful for PLS or other models that expect 1D input.

  • weighted_sum: Computes a weighted combination of channel values at each location. Useful when channels should be fused into a single signal.

  • mean: Simple average across channels (special case of weighted_sum).

Note

Channels must be aligned to the same grid before combining. Use step_measure_channel_align() first if grids differ.

See also

Other measure-channel: step_measure_channel_align(), step_measure_channel_ratio()

Examples

library(recipes)
library(tibble)

# Create sample multi-channel data (already aligned)
df <- tibble(
  id = rep(1:3, each = 10),
  time = rep(seq(0, 9, by = 1), 3),
  uv = rnorm(30, 100, 10),
  ri = rnorm(30, 50, 5),
  concentration = rep(c(10, 25, 50), each = 10)
)

# Ingest and combine with stacking
rec <- recipe(concentration ~ ., data = df) |>
  update_role(id, new_role = "id") |>
  step_measure_input_long(uv, location = vars(time)) |>
  step_measure_input_long(ri, location = vars(time)) |>
  step_measure_channel_combine(strategy = "stack")