step_nnmf_sparse: Non-Negative Matrix Factorization Signal Extraction with...

In recipes: Preprocessing and Feature Engineering Steps for Modeling

Non-Negative Matrix Factorization Signal Extraction with lasso Penalization

Description

step_nnmf_sparse() creates a specification of a recipe step that will convert numeric data into one or more non-negative components.

Usage

step_nnmf_sparse(
  recipe,
  ...,
  role = "predictor",
  trained = FALSE,
  num_comp = 2,
  penalty = 0.001,
  options = list(),
  res = NULL,
  prefix = "NNMF",
  seed = sample.int(10^5, 1),
  keep_original_cols = FALSE,
  skip = FALSE,
  id = rand_id("nnmf_sparse")
)

Arguments

recipe	A recipe object. The step will be added to the sequence of operations for this recipe.
...	One or more selector functions to choose variables for this step. See selections() for more details.
role	For model terms created by this step, what analysis role should they be assigned? By default, the new columns created by this step from the original variables will be used as predictors in a model.
trained	A logical to indicate if the quantities for preprocessing have been estimated.
num_comp	The number of components to retain as new predictors. If num_comp is greater than the number of columns or the number of possible components, a smaller value will be used. If num_comp = 0 is set then no transformation is done and selected variables will stay unchanged, regardless of the value of keep_original_cols.
penalty	A non-negative number used as a penalization factor for the loadings. Values are usually between zero and one.
options	A list of options to nmf() in the RcppML package. That package has a separate function setRcppMLthreads() that controls the amount of internal parallelization. Note that the argument A, k, L1, and seed should not be passed here.
res	A matrix of loadings is stored here, along with the names of the original predictors, once this preprocessing step has been trained by prep().
prefix	A character string for the prefix of the resulting new variables. See notes below.
seed	An integer that will be used to set the seed in isolation when computing the factorization.
keep_original_cols	A logical to keep the original variables in the output. Defaults to FALSE.
skip	A logical. Should the step be skipped when the recipe is baked by bake()? While all operations are baked when prep() is run, some operations may not be able to be conducted on new data (e.g. processing the outcome variable(s)). Care should be taken when using skip = TRUE as it may affect the computations for subsequent operations.
id	A character string that is unique to this step to identify it.

Details

Non-negative matrix factorization computes latent components that have non-negative values and take into account that the original data have non-negative values.

The argument num_comp controls the number of components that will be retained (the original variables that are used to derive the components are removed from the data). The new components will have names that begin with prefix and a sequence of numbers. The variable names are padded with zeros. For example, if num_comp < 10, their names will be NNMF1 - NNMF9. If num_comp = 101, the names would be NNMF1 - NNMF101.

Value

An updated version of recipe with the new step added to the sequence of any existing operations.

Tidying

When you tidy() this step, a tibble with column terms (the selectors or variables selected) and the number of components is returned.

Tuning Parameters

This step has 2 tuning parameters:

• num_comp: # Components (type: integer, default: 2)

• penalty: Amount of Regularization (type: double, default: 0.001)

Case weights

The underlying operation does not allow for case weights.

See Also

Other multivariate transformation steps: step_classdist_shrunken(), step_classdist(), step_depth(), step_geodist(), step_ica(), step_isomap(), step_kpca_poly(), step_kpca_rbf(), step_kpca(), step_mutate_at(), step_nnmf(), step_pca(), step_pls(), step_ratio(), step_spatialsign()

Examples

if (rlang::is_installed(c("modeldata", "RcppML", "ggplot2"))) {
library(Matrix)
data(biomass, package = "modeldata")

rec <- recipe(HHV ~ ., data = biomass) %>%
  update_role(sample, new_role = "id var") %>%
  update_role(dataset, new_role = "split variable") %>%
  step_nnmf_sparse(
    all_numeric_predictors(),
    num_comp = 2,
    seed = 473,
    penalty = 0.01
  ) %>%
  prep(training = biomass)

bake(rec, new_data = NULL)

library(ggplot2)
bake(rec, new_data = NULL) %>%
  ggplot(aes(x = NNMF2, y = NNMF1, col = HHV)) +
  geom_point()
}

recipes documentation built on Aug. 26, 2023, 1:08 a.m.