classify: Classification of the one dimensional points in a Pcp or Mlp...

Description Usage Arguments Details Value Author(s) Examples

Description

Classification based on ROC params (TN TP FP FN).

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
## S4 method for signature 'ClassifiedPoints'
getData(x, n = NULL)

classify(x, ...)

## S4 method for signature 'Pcp'
classify(x, ...)

## S4 method for signature 'Mlp'
classify(x, ...)

## S4 method for signature 'ClassifiedPoints'
initialize(.Object, ..., scores,
  scores.points = scores.points, scores.index = scores.index, ROC, AUC,
  class.color)

## S4 method for signature 'ClassifiedPoints,missing'
plot(x, y, comparison = "all",
  class.color = NULL, ...)

## S4 method for signature 'ClassifiedPoints'
show(object)

Arguments

x

Pcp or Mlp Object for the function classify otherwise it is a ClassifiedPoints object

n

data to extract from ClassifiedPoints (NULL gives all)

...

additional arguments to pass on

.Object

internal object

scores

final scores

scores.points

sorted points

scores.index

index of sorted points

ROC

parameters (TN, TP, FN and FP)

AUC

area under the curve

class.color

user assigned group coloring scheme

y

default plot param, which should be set to NULL

comparison

Specify a comparison i.e. ("grp1 vs grp2") and plot only that comparison.

object

ClassifiedPoints Object

Details

Tests all possible discrimination lines and picks the one with highest score based on a score which is simply calculated by the formula (TP - FP) + (TN - FN).

The plot shows the distribution of scores for different discrimination lines. Each line is a separator that has a score for the separation of the two groups, and the height of the line marks the score for this separation.

Value

The classify function returns an object of class ClassifiedPoints

Author(s)

Jesper R. Gadin and Jason T. Serviss

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
#use demo data
data(pcpMatrix)
classes <- rownames(pcpMatrix)

#run function
prj <- pcp(pcpMatrix, classes)
cl <- classify(prj)

#getData accessor
getData(cl)

#getData accessor specific
getData(cl, "scores")

#plot result
plot(cl)

ClusterSignificance documentation built on Nov. 8, 2020, 5:28 p.m.