-1
$\begingroup$

This is the code for k-seeds clustering where features and side_effects are the excel files. I am facing problem in understanding the results. Where to find the calculated AUC and AUPR values? And what does this implies (list(AUCFinal,AUPRFinal)) where, first value is the mean AUC on the various folders, second value is the mean AUPR on the various folders. the code is -

#vectors where the means of the AUC will be stored
vectorMedieAUC<-numeric()
#vectors where the means of the AUPR will be stored
vectorMedieAUPR<-numeric()

#various Iterations of the k-folds cross validation procedure
for(j in 1:num_iterations){
  i=0 # variable to trace that is the first time of the k-fold cross validation procedure
  vectorAUC<-numeric() #instantiate a vecotr of AUC for first iteration of the k-fold cross validation procedure
  vectorAUPR<-numeric() # vector of AUPR for first iteration of the k-folds cross validation procedure

  folds<-CreateFolds(features,num_folds)
  train = features[folds != i,]
  trainpharmat = side_effects[folds != i,]
  test = features[folds == i,]
  testpharmat = side_effects[folds == i,]

  #KSeeds clustering method application
  s<-RandomSeedGenerator(num_clusters,nrow(train)) #function that generates randomly the numbers that will be the seeds of the cluster
  Seed<-SeedSelection(train,num_clusters,s)
  #Return the list of clusters
  clusters<-KSeedsClusters (train,num_clusters,Seed,s)
  #Return the matrix A of KSeeds scores
  A<-KSeedsScores(train,trainpharmat,num_clusters,Seed,s,clusters)

  #KSeeds predictions for drugs in the test set
  predizioni<-PredictionKSeeds(test,Seed,num_clusters,A)

  #Function for obtaining AUC
  vectorAUC<-AUC(predizioni,testpharmat,vectorAUC,"KSeeds")

  #Function for obtaining AUPR
  vectorAUPR<-AUPR(predizioni,testpharmat,vectorAUPR,"KSeeds")

  for(i in 1:(num_folds-1)){
    train = features[folds != i,]
    trainpharmat = side_effects[folds != i,]
    test = features[folds == i,]
    testpharmat = side_effects[folds == i,]

    #KSeeds clustering
    s<-RandomSeedGenerator(num_clusters,nrow(train)) #function that generates randomly the numbers that will be the seeds of the cluster
    Seed<-SeedSelection(train,num_clusters,s)

    #Return the list of clusters
    clusters<-KSeedsClusters (train,num_clusters,Seed,s)
    #Return the matrix A of KSeeds scores
    A<-KSeedsScores(train,trainpharmat,num_clusters,Seed,s,clusters)





    #KSeeds predictions for drugs in the test set
    predizioni<-PredictionKSeeds(test,Seed,num_clusters,A)

    #Function for obtaining AUC
    vectorAUC<-AUC(predizioni,testpharmat,vectorAUC,"KSeeds")

    #Function for obtaining AUPR
    vectorAUPR<-AUPR(predizioni,testpharmat,vectorAUPR,"KSeeds")
  }
  vectorMedieAUC<-c(vectorMedieAUC,mean(vectorAUC))
  vectorMedieAUPR <-c(vectorMedieAUC,mean(vectorAUPR))
  j=j+1

}

AUCFinal<-mean(vectorMedieAUC)
AUPRFinal<-mean(vectorMedieAUPR)
return(list(AUCFinal,AUPRFinal))

}
$\endgroup$
0
$\begingroup$

The calculated values are what you compute the means from:

AUCFinal<-mean(vectorMedieAUC)
AUPRFinal<-mean(vectorMedieAUPR)

But somehow I have the impression that you googled some code that you do not understand... But then, on the other hand, it's from a pretty dead R package that like many other R packages breaks with most programming conventions...

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.