R/svm.functions.R

In OmicsMarkeR: Classification and Feature Selection for 'Omics' Datasets

#' @title SVM Recursive Feature Extraction (Binary)
#' @description This conducts feature selection for Support Vector Machines 
#' models via recursive feature extraction.  This returns a vector of the 
#' features in x ordered by relevance. The first item of the vector has the 
#' index of the feature which is more relevant to perform the classification 
#' and the last item of the vector has the feature which is less relevant.
#' This function is specific to Binary classification problems,
#' @param x A matrix where each column represents a feature and each row 
#' represents a sample
#' @param y A vector of labels corresponding to each sample's group membership
#' @param c A numeric value corresponding to the 'cost' applied during the 
#' svm model fitting. This can be selected by the user if using this 
#' function directly or is done internally.
#' @param perc.rem A numeric value indicating the percent of features 
#' removed during each iteration. Default \code{perc.rem = 10}.
#' @return Vector of features ranked from most important to least important.
#' @references Guyon I. et. al. (2010) \emph{Gene Selection for Cancer 
#' Classification using Support Vector Machines}. Machine Learning 46 389-422. 
#' 
#' @seealso \code{\link{svmrfeFeatureRankingForMulticlass}}
#' @example inst/examples/svmrfeFeatureRanking.R 
#' @import e1071
#' @export

svmrfeFeatureRanking = function(x, y, c, perc.rem=10)
{
    n = ncol(x)
    survivingFeaturesIndexes = seq(n)
    featureRankedList = vector(length=n)
    rankedFeatureIndex = n
    while(length(survivingFeaturesIndexes)>0){
        #train the support vector machine
        svmModel <- svm(x[, survivingFeaturesIndexes], 
                        y, 
                        cost = c, 
                        cachesize=500,
                        scale= FALSE, 
                        type="C-classification", kernel="linear" )
        #compute the weight vector
        w <- t(svmModel$coefs)%*%svmModel$SV
        #compute ranking criteria
        rankingCriteria <- w * w
        #rank the features
        ranking <- sort(rankingCriteria, index.return = TRUE)$ix
        
        ## New to determine 10% of remaining features
        #round to make sure a defined number of features to remove 
        # (ceiling rounds up)
        r <- ceiling(length(survivingFeaturesIndexes)/perc.rem)
        s <- length(survivingFeaturesIndexes)
        
        ## New update feature ranked list
        featureRankedList[rev((s-r+1):s)] <- 
            survivingFeaturesIndexes[ranking[1:r]]
        
        ## New to remove 10%
        rankedFeatureIndex <- rankedFeatureIndex - r
        
        #eliminate the feature with smallest ranking criterion
        survivingFeaturesIndexes <- 
            survivingFeaturesIndexes[-ranking[1:r]]
    }
    return (featureRankedList)
}

#' @title SVM Multiclass Weights Ranking
#' @description This calculates feature weights for multiclass Support 
#' Vector Machine (SVM) problems
#' @param model A fitted SVM model of multiclass
#' @return Vector of feature weights
#' @references Guyon I. et. al. (2010) \emph{Gene Selection for Cancer 
#' Classification using Support Vector Machines}. Machine Learning 46 389-422. 

svm.weights<-function(model){
    w=0
    if(model$nclasses==2){
        w=t(model$coefs)%*%model$SV
    }else{ #when we deal with OVO svm classification
        ## compute start-index
        start <- c(1, cumsum(model$nSV)+1)
        start <- start[-length(start)]
        calcw <- function (i,j) {
            ## ranges for class i and j:
            ri <- start[i] : (start[i] + model$nSV[i] - 1)
            rj <- start[j] : (start[j] + model$nSV[j] - 1)
            ## coefs for (i,j):
            coef1 <- model$coefs[ri, j-1]
            coef2 <- model$coefs[rj, i]
            ## return w values:
            w=t(coef1)%*%model$SV[ri,]+t(coef2)%*%model$SV[rj,]
            return(w)
        }
        W=NULL
        for (i in 1 : (model$nclasses - 1)){
            for (j in (i + 1) : model$nclasses){
                wi=calcw(i,j)
                W=rbind(W,wi)
            }
        }
        w=W
    }
    return(w)
}

#' @title SVM Recursive Feature Extraction (Multiclass)
#' @description This conducts feature selection for Support Vector Machines 
#' models via recursive feature extraction.  This returns a vector of the 
#' features in x ordered by relevance. The first item of the vector has the 
#' index of the feature which is more relevant to perform the classification 
#' and the last item of the vector has the feature which is less relevant.
#' This function is specific to Binary classification problems.
#' @param x A matrix where each column represents a feature and each row 
#' represents a sample
#' @param y A vector of labels corresponding to each sample's group membership
#' @param c A numeric value corresponding to the 'cost' applied during the 
#' svm model fitting. This can be selected by the user if using this function 
#' directly or is done internally.
#' @param perc.rem A numeric value indicating the percent of features removed 
#' during each iteration. Default \code{perc.rem = 10}.
#' @return Vector of features ranked from most important to least important.
#' @references Guyon I. et. al. (2010) \emph{Gene Selection for Cancer 
#' Classification using Support Vector Machines}. Machine Learning 46 389-422. 
#' @seealso \code{\link{svmrfeFeatureRanking}}
#' @example inst/examples/svmrfeFeatureRankingForMulticlass.R
#' @export

svmrfeFeatureRankingForMulticlass = function(x,y,c, perc.rem = 10){
    n = ncol(x)
    survivingFeaturesIndexes = seq(n)
    featureRankedList = vector(length=n)
    rankedFeatureIndex = n
    while(length(survivingFeaturesIndexes)>0){
        #train the support vector machine
        svmModel = svm(x[, survivingFeaturesIndexes], 
                       y, 
                       cost = c, 
                       cachesize=500,
                       scale= FALSE, 
                       type="C-classification", kernel="linear" )
        #compute the weight vector
        multiclassWeights = svm.weights(svmModel)
        #compute ranking criteria
        multiclassWeights = multiclassWeights * multiclassWeights
        rankingCriteria = 0
        for(i in 1:ncol(multiclassWeights))rankingCriteria[i] =
            mean(multiclassWeights[,i])
        #rank the features
        (ranking = sort(rankingCriteria, index.return = TRUE)$ix)
        
        ## New to determine prec.rem of remaining features
        #round to make sure a defined number of features to remove 
        # ceiling rounds up)
        r <- ceiling(length(survivingFeaturesIndexes)/perc.rem)
        s <- length(survivingFeaturesIndexes)
        
        ## New update feature ranked list
        featureRankedList[rev((s-r+1):s)] <- 
            survivingFeaturesIndexes[ranking[1:r]]
        
        ## New to remove perc.rem
        rankedFeatureIndex <- rankedFeatureIndex - r
        
        #eliminate the feature with smallest ranking criterion
        survivingFeaturesIndexes <- 
            survivingFeaturesIndexes[-ranking[1:r]]
    }
    return(featureRankedList)
}