Universal fit function (Parallel)
fit.parallel(data, dep, indep, classifier = "lr", classifier.params = list(rf.ntree = 100, rf.mtry = NULL, c5.0.trials = 40, c5.0.rules = TRUE, c5.0.winnow = FALSE, nb.fL = 0, nb.adjust = 1, svm.gamma = NULL, svm.cost = 1), params.tuning = FALSE, normalize = "no", rebalance = "no", validation = "boot", validation.params = list(cv.k = 10, boot.n = 100), prob.threshold = 0.5, repeats = 1, n.cores = 2)
| data | a dataframe for input data |
|---|---|
| dep | a character for dependent variable |
| indep | a vector of characters for independent variables |
| classifier | a character for classifier techniques, i.e., lr, rf, c5.0, nb, and svm |
| classifier.params | a list of parameters for an input classifier technique |
| params.tuning | a boolean indicates whether to perform parameters tuning |
| normalize | a character for normalization techniques, i.e., log, scale, center, standardize, and no for non-normalization#' |
| rebalance | a character for a choice of data sampling techniques, i.e., up for upsampling, down for downsampling, and no for no-sampling (default: "NO") |
| validation | a character for a choice of validation techniques, i.e., boot for bootstrap validation technique, cv for cross-validation technique, and no for constructing a model with the whole dataset without model validation (default: "boot") |
| validation.params | a list of parameters for an input validation techniques (default: list(cv.k = 10, boot.n = 100)) |
| prob.threshold | a numeric for probability threshold (default: 0.5) |
| repeats | a numeric for number of repetitions (default: 1) |
| n.cores | a numeric for number of cores (default: 2) |