elm: Extreme learning machine for survival, classification,...

In linxihui/sml: Machine Learning for Survival Data

Description

Extreme learning machine for survival, classification, regression (univariate / multivariate)

Usage

elm(x, ...)

## S3 method for class 'formula'
elm(formula, data = environment(formula), ...)

## Default S3 method:
elm(x, y, nhid = max(200, 2 * ncol(x)), actfun = "sig",
  scaled = TRUE, family = switch(class(y), Surv = "cox", factor =
  ifelse(nlevels(y) > 2, "multinomial", "binomial"), integer = "poisson", matrix
  = "mgaussian", "gaussian"), lambda = ifelse(family == "mgaussian", 0, 1),
  ...)

## S3 method for class 'elm'
predict(object, newdata, type = c("response", "probabilities",
  "link", "risk"))

Arguments

x	Design matrix
...	Further parameters passed to internal function
formula	Model formula
data	Training data frame
y	Survival object, factor or a vector/matrix of continuous variable
nhid	Number of (random) hidden neurons
actfun	Activating function
scaled	If to standardize input units
family	Choices of 'cox', 'binomial', 'multinomial', 'poisson', 'gaussian', 'mgaussian'
lambda	Shrinkage factor
object	elm object
newdata	A data frame (if elm called via formula) or a design matrix (if elm called via design matrix)
type	Type of output

Value

elm object.

Examples

library(survival);
data(pbc, package = 'randomForestSRC');
pbc <- na.omit(pbc);
i.tr <- sample(nrow(pbc), 100);

elm.f <- elm(Surv(days, status) ~., data = pbc[i.tr, ], nhid = 500);
elm.pred <- predict(object = elm.f, newdata = pbc[-i.tr, ], type = 'link');
survConcordance(Surv(days, status) ~ elm.pred, data = pbc[-i.tr, ])

linxihui/sml documentation built on May 21, 2019, 6:39 a.m.