instance_selection: Instance selection wrapper

In smartdata: Data Preprocessing

Description

Instance selection wrapper

Usage

instance_selection(dataset, method, class_attr = "Class", ...)

Arguments

dataset	we want to perform an instance selection on
method	selected method of instance selection
class_attr	character. Indicates the class attribute from dataset. Must exist in it
...	Further arguments for method

Value

A filtered dataset with same shape as the input to the function

Examples

library("smartdata")

super_iris <- instance_selection(iris, method = "CNN", class_attr = "Species")
# Use k = 2 instead of default k
super_iris <- instance_selection(iris, method = "CNN", class_attr = "Species", k = 2)
# Use Edited Nearest Neighbor as method to select observations
super_iris <- instance_selection(iris, method = "ENN", class_attr = "Species", k = 3)
super_iris <- instance_selection(iris, method = "multiedit", class_attr = "Species",
                   k = 3, num_folds = 5, null_passes = 8)
# Use default arguments for multiedit
super_iris <- instance_selection(iris, method = "multiedit", class_attr = "Species")
super_iris <- instance_selection(iris, method = "FRIS", class_attr = "Species")
# FRIS method with fuzzy granularity of 2
super_iris <- instance_selection(iris, method = "FRIS", class_attr = "Species", alpha = 2)
# FRIS method with Dubois Prade implicator
super_iris <- instance_selection(iris, method = "FRIS", "Species", implicator_type = "dubois_prade")
# FRIS method with lower threshold (that is, less observations are removed)
super_iris <- instance_selection(iris, method = "FRIS", class_attr = "Species", threshold = 0.6)

smartdata documentation built on Dec. 19, 2019, 1:08 a.m.