r.optimize.threshold: r.optimize.threshold

In rocalabern/rmodel: Tools for data analysis

Usage

r.optimize.threshold(score, target, f_obj = function(threshold) {     predict
  = numeric(length(score))     predict[score > threshold] = 1     t <-
  table(predict, target)     if (nrow(t) == 1)          f1 = 0     else f1 = 0.5
  * (2 * t[1, 1]/(2 * t[1, 1] + t[1, 2] + t[2, 1])) + 0.5 * (2 * t[2, 2]/(2 *
  t[2, 2] + t[1, 2] + t[2, 1]))     return(-f1) }, x0 = quantile(score, probs
  = 0.5), xtol_rel = 1e-08, maxtime = 60, algorithm = "NLOPT_LN_COBYLA")

rocalabern/rmodel documentation built on May 27, 2019, 12:14 p.m.