This process is aimed to remove noisy, superuous, or inconsistent instances from training datasets but retain consistent ones. Therefore, good accuracy of classi cation is achieved by removing instances which do not give positive contributions. In the package, we provide the following functions:

  •  IS.FRIS.FRST: This is a function that implements the fuzzy-rough in stance selection (FRIS) proposed by (R. Jensen and C. Cornelis, 2010) which is used to perform instance selection.
  •  IS.FRPS.FRST: This is a function for implementing instance selection using prototype selection method (FRPS) proposed by (N. Verbiest et al,2013).

It should be noted that outputs of these functions are indexes of selected instances. So, to generate a new decision table we need to execute the SF.applyDecTable function.

 

  • IS.FRIS.FRST

R> #############################################
R> ## Example: Evaluate instances/objects and
R> ## generate new decision table
R> #############################################
R> dt.ex1 <- data.frame(c(0.1, 0.5, 0.2, 0.3, 0.2, 0.2, 0.8),
+ c(0.1, 0.1, 0.4, 0.2, 0.4, 0.4, 0.5), c(0, 0, 0, 0, 1, 1, 1))
R> colnames(dt.ex1) <- c("a1", "a2", "d")
R> decision.table <- SF.asDecisionTable(dataset = dt.ex1, decision.attr = 3, indx.nominal = R> ## evaluate index of objects
R> res.1 <- IS.FRIS.FRST(decision.table = decision.table, control =
+ list(threshold.tau = 0.5, alpha = 1,
+ type.aggregation = c("t.tnorm", "lukasiewicz"),
+ t.implicator = "lukasiewicz"))
R> print(res.1)
$indx.objects
[1] 2 7
$type.method
[1] "IS.FRIS"
$type.task

[1] "instance selection"
$type.model
[1] "FRST"
attr(,"class")
[1] "InstanceSelection" "list"

R> ## generate new decision table
R> new.decTable <- SF.applyDecTable(decision.table, res.1)

 

  • IS.FRPS.FRST

R> #############################################
R> ## Example: Evaluate instances/objects and
R> ## generate new decision table
R> #############################################
R> dt.ex1 <- data.frame(c(0.5, 0.2, 0.3, 0.7, 0.2, 0.2), c(0.1, 0.4, 0.2, 0.8, 0.4, 0.4), c(0, 0, 0, 1, 1, 1))

R> colnames(dt.ex1) <- c("a1", "a2", "d")
R> decision.table <- SF.asDecisionTable(dataset = dt.ex1, decision.attr = 3,indx.nominal = c(3))

R> ## evaluate instances
R> res.1 <- IS.FRPS.FRST(decision.table, type.alpha = "FRPS.3")
R> print(res.1)
$indx.objects
[1] 1 3 4
$type.method
[1] "IS.FRPS"
$type.task
[1] "instance selection"
$type.model
[1] "FRST"
attr(,"class")
[1] "InstanceSelection" "list"


R> ## generate new decision table
R> new.decTable <- SF.applyDecTable(decision.table, res.1)