find.optimal.thresh.cv: Function to find the optimal RANKS score thereshold

Description

Function to find the optimal quantile alpha and corresponding threshold by cross-validation with a kernel-based score method.

Usage

find.optimal.thresh.cv(K, ind.pos, ind.non.pos, m = 5, 
alpha = seq(from = 0.05, to = 0.6, by = 0.05), init.seed = NULL, 
opt.fun = compute.F, fun = KNN.score, ...)

Value

A list with 3 elements:

alpha: quantile corresponding to the best F-score
thresh: threshold corresponding to the best F-score
pos.scores: scores of the positive elements computed through CV

Arguments

K

matrix. Kernel matrix or any valid symmetric matrix

ind.pos

indices of the positive examples. They are the indices the row of RW corresponding to positive examples of the training set.

ind.non.pos

indices of the non positive examples. They are the indices the row of RW corresponding to non positive examples of the training set.

m

number of folds (default: 5)

alpha

vector of the quantiles to be tested

init.seed

initial seed for the random generator. If NULL (def) no initialization is performed

opt.fun

Function implementing the metric to select the optimal threshold. The F-score (compute.F) is the default. Available functions:

- compute.F: F-score (default)

- compute.acc:accuracy.

Any function having two arguments representing the vector of predicted and true labels can be in principle used.

fun

function. It must be a kernel-based score method (default KNN.score)

...

optional arguments for the function fun

Details

Function to find the optimal quantile alpha and corresponding threshold by cross-validation with a kernel-based score method. The optimality is computed with respect to a specific metric (def: F-score). This function is used by multiple.ker.score.thresh.cv, ker.score.classifier.holdout, ker.score.classifier.cv.

Examples

Run this code

# Finding the optimal threshold in the Tanimoto chemical structure similarity network 
# between 1253 DrugBank drugs for the prediction of the DrugBank category Penicillins using
# the KNN-score with the random walk kernel 
library(bionetdata);
data(DD.chem.data);
data(DrugBank.Cat);
K <- rw.kernel(DD.chem.data);
labels <- DrugBank.Cat[,"Penicillins"];
ind.pos <- which(labels==1);
ind.non.pos <- which(labels==0);
res <- find.optimal.thresh.cv(K, ind.pos, ind.non.pos);
res