Predict Method for Feature Selection SVM

This function predicts values based upon a model trained by svm. If class assigment is provided, confusion table, missclassification table, sensitivity and specificity are calculated.

Usage

# S3 method for class 'penSVM'
predict(object, newdata, newdata.labels = NULL, labels.universe=NULL, ...)

Arguments

object

Object of class "penSVM", created by 'svmfs'

newdata

A matrix containing the new input data, samples in rows, features in columns

newdata.labels

optional, new data class labels

labels.universe

important for models produced by loocv: all possible labels in the particular data set

...

additional argument(s)

Value

returns a list of prediction values for classes

pred.class

predicted class

tab

confusion table

error

missclassification error

sensitivity

sensitivity

specificity

specificity

Author

Natalia Becker

See also

svm, svmfs

Examples

seed<- 123
train<-sim.data(n = 200, ng = 100, nsg = 10, corr=FALSE, seed=seed )
print(str(train)) 
#> List of 3
#>  $ x   : num [1:100, 1:200] 0.547 0.635 -0.894 0.786 2.028 ...
#>   ..- attr(*, "dimnames")=List of 2
#>   .. ..$ : chr [1:100] "pos1" "pos2" "pos3" "pos4" ...
#>   .. ..$ : chr [1:200] "1" "2" "3" "4" ...
#>  $ y   : Named num [1:200] 1 -1 -1 1 -1 -1 1 -1 -1 -1 ...
#>   ..- attr(*, "names")= chr [1:200] "1" "2" "3" "4" ...
#>  $ seed: num 123
#> NULL

#train standard svm
my.svm<-svm(x=t(train$x), y=train$y, kernel="linear")

# test with other data
test<- sim.data(n = 200, ng = 100, nsg = 10, seed=(seed+1) )

# Check accuracy standard SVM
my.pred <-ifelse( predict(my.svm, t(test$x)) >0,1,-1)
# Check accuracy:
table(my.pred, test$y)
#>        
#> my.pred -1  1
#>      -1 81 23
#>      1  30 66

if (FALSE) # define set values of tuning parameter lambda1 for SCAD 
lambda1.scad <- c (seq(0.01 ,0.05, .01),  seq(0.1,0.5, 0.2), 1 ) 
# for presentation don't check  all lambdas : time consuming! 
# computation intensive; for demostration reasons only for the first 100 features
# and only for 10 Iterations maxIter=10, default maxIter=700

system.time(fit.scad<- svmfs(x=t(train$x)[,1:100],y=train$y, fs.method="scad", cross.outer= 0,
  grid.search = "discrete",  lambda1.set=lambda1.scad[1:3], show="none",
  parms.coding = "none", maxIter=10,
  inner.val.method = "cv", cross.inner= 5, seed=seed, verbose=FALSE))
#> [1] "grid search"
#> [1] "discrete"
#> [1] "show"
#> [1] "none"
#> [1] "feature selection method is scad"
#> Error: object 'lambda1.scad' not found
#> Timing stopped at: 0.001 0.001 0.001


# SCAD 
test.error.scad<-predict(fit.scad, newdata=t(test$x)[,1:100],newdata.labels=test$y )
#> Error: object 'fit.scad' not found
 # Check accuracy SCAD SVM    
print(test.error.scad$tab)  
#> Error: object 'test.error.scad' not found
 # \dontrun{}

#########################################
## analog for 1-norm SVM
#epsi.set<-vector(); for (num in (1:9)) epsi.set<-sort(c(epsi.set, c(num*10^seq(-5, -1, 1 ))) )
#lambda1.1norm <-   epsi.set[c(3,5)] # 2 params
#
## train 1norm SVM
# norm1.fix<- svmfs(t(train$x), y=train$y, fs.method="1norm", 
#    cross.outer= 0, grid.search = "discrete",  
#    lambda1.set=lambda1.1norm, show="none",
#    parms.coding = "none",
#    maxIter = 700, inner.val.method = "cv", cross.inner= 5,
#    seed=seed, verbose=FALSE )   
#  
#  print(norm1.fix)   
#
## L1-norm SVM
#test.error.1norm<-predict(norm1.fix, newdata=t(test$x),newdata.labels=test$y )
# # Check accuracy L1-norm SVM    
#print(test.error.1norm$tab)