ModelSelectionResultAccessors: ModelSelectionResult Accessors

Description Usage Arguments Value Accessor-like methods Author(s) References Examples

Description

ModelSelectionResult Accessors

Usage

1
2
## S4 method for signature 'ModelSelectionResult'
gridRows(object)

Arguments

object

a model selection result object (can be extracted from KeBABS model with accessor modelSelResult)

Value

gridRows: returns a list of kernel objects
gridColumns: returns a DataFrame object with grid column parameters
gridErrors: returns a matrix with grid errors
performance: returns a list of matrices with performance values selGridRow: returns the selected kernel selGridCol: returns the selected SVM and/or hyperparameter(s) fullModel: returns a kebabs model of class KBModel

Accessor-like methods

gridRows: return the grid rows containing the kernels.

gridColumns: return the grid columns.

gridErrors: return the grid CV errors.

performance: return the collected performance parameters.

selGridRow: return the selected grid row.

selGridCol: return the selected grid column.

fullModel: return the full model.

Author(s)

Johannes Palme <[email protected]>

References

http://www.bioinf.jku.at/software/kebabs

J. Palme, S. Hochreiter, and U. Bodenhofer (2015) KeBABS: an R package for kernel-based analysis of biological sequences. Bioinformatics, 31(15):2574-2576, 2015. DOI: 10.1093/bioinformatics/btv176.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
## create kernel object for normalized spectrum kernel
specK5 <- spectrumKernel(k=5)
## Not run: 
## load data
data(TFBS)

## perform training - feature weights are computed by default
model <- kbsvm(enhancerFB, yFB, specK5, pkg="LiblineaR",
               svm="C-svc", cost=c(1,15,50,100), cross=10,
               perfParameters="ALL", showProgress=TRUE)

## show model selection result
mres <- modelSelResult(model)
mres

## extract grid errors
gridErrors(mres)

## extract other performance parameters
performance(mres)

## End(Not run)

kebabs documentation built on Nov. 17, 2017, 1:20 p.m.