DaMiR.FBest: Select best predictors to build Classification Model

In BioinfoMonzino/DaMiRseq: Data Mining for RNA-seq data: normalization, feature selection and classification

Description

This function allows the user to select a subset of predictors; the number of predictors can be defined by user or selected automatically.

Usage

DaMiR.FBest(
  data,
  ranking,
  autoselect = c("no", "yes"),
  n.pred = 10,
  th.zscore = 2
)

Arguments

data	A transposed data frame of expression data. Rows and Cols should be, respectively, observations and features
ranking	A data frame with importance score for each feature, generated by DaMiR.FSort
autoselect	A flag to specify how to select predictors: "no" (default) - Manually: users can specify the number of best predictors, setting n.pred argument "yes" - Automatically: users have to specify the importance threshold defined by the th.zscore argument; features will be accordingly selected
n.pred	If autoselect="no" then the user have to specify the number of predictors; default is 10
th.zscore	Threshold of scaled importance score (Z-score); default value is 2

Value

A list containing:

• A data frame of normalized expression data of the most important selected predictors.

• A vector with predictors name.

Author(s)

Mattia Chiesa, Luca Piacentini

See Also

DaMiR.FSort

Examples

# use example data:
data(data_reduced)
data(data_relief)
# select the first 8 predictors rankad by imporatance:
selected_features <- DaMiR.FBest(data_reduced, data_relief, n.pred = 8)
# select predictors by importance but automatically:
selected_features <- DaMiR.FBest(data_reduced, data_relief,
autoselect = "yes", th.zscore = 1.5)

BioinfoMonzino/DaMiRseq documentation built on Aug. 22, 2021, 3:11 p.m.