These internal functions are not intended for direct use by users.
For example, err obtains error messages from Fortran code,
error.bars helps to plot cross-validation error curves,
and getmin identifies the best lambda through cross-validation.
Other functions assist with kernel loss computations, coefficient management, and interpolation.
cv.gaussian(outlist, lambda, x, y, foldid, pred.loss, hval)
cv.uniform(outlist, lambda, x, y, foldid, pred.loss, hval)
cv.epanechnikov(outlist, lambda, x, y, foldid, pred.loss, hval)
cvcompute(mat, foldid, nlams)
dcsvm.gauss(x, y, alpha, lam2, hval, nlam, flmin, ulam, isd, eps, dfmax, pmax, jd,
pfncol, pf, pf2, maxit, istrong, nobs, nvars, vnames)
dcsvm.unif(x, y, alpha, lam2, hval, nlam, flmin, ulam, isd, eps, dfmax, pmax, jd,
pfncol, pf, pf2, maxit, istrong, nobs, nvars, vnames)
dcsvm.epane(x, y, alpha, lam2, hval, nlam, flmin, ulam, isd, eps, dfmax, pmax, jd,
pfncol, pf, pf2, maxit, istrong, nobs, nvars, vnames)
err(n, maxit, pmax)
error.bars(x, upper, lower, width = 0.02, ...)
getmin(lambda, cvm, cvsd)
getoutput(fit, maxit, pmax, nvars, vnames)
lambda.interp(lambda, s)
lamfix(lam)
loss.epanechnikov(u, hval)
loss.gaussian(u, hval)
loss.uniform(u, hval)
nonzero(beta, bystep = FALSE)
zeromat(nvars, nalam, vnames, stepnames)Internal DCSVM Functions
Internal helper functions for the DCSVM package.
Most of these functions are adapted or directly copied from the gcdnet and glmnet packages.