Description Usage Arguments Value Note Author(s) References Examples
Fit Extreme Learning Machine.
1 2 3 |
X.fit |
Data matrix (numeric) containing the input values (predictors) used to train the model. |
Y.fit |
Response vector (numeric) used to train the model. |
Number.hn |
Number of units in the hidden layer. Default is |
autorangeweight |
Option whether to use the automated range used for the weights. Default is |
rangeweight |
Initial random weights on |
activation |
Activation function of the hidden layer neurons. Available functions are: 'TANH' (default) and 'SIG'. |
outputBias |
Option whether to use the bias parameter in the output layer |
rangebias |
Initial random bias on |
inputWeight |
Set of weights used. |
biasofHN |
Set of bias used. |
matrixBeta |
Set of weights adjusted. |
matrixP |
matrixP. |
predictionTrain |
The fitted values for the training data. |
rangeweight |
Used range of the random weight initialization ( |
.
activation |
Activation function of the hidden layer neurons. |
outputBias |
Option whether to use the bias parameter in the output layer. |
rangebias |
Used range of the random bias initialization ( |
To achieve better results, the use of a pre-processing technique (e.g. standardization of variables) is important.
Aranildo Lima
G.-B. Huang, Q.-Y. Zhu, C.-K. Siew (2006) Extreme learning machine: Theory and applications Neurocomputing 70 (2006) 489-501
Lima, A.R.; A.J. Cannon and W.W. Hsieh. Nonlinear Regression In Environmental Sciences Using Extreme Learning Machines. Submited to: Environmental Modelling and Software - ELSEVIER (submitted 2014/2/3).
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 | set.seed(123)
attach(wtloss)
library("scales")
#scaling the inputs/outputs
x.train <- rescale(as.matrix(wtloss$Days), to=c(-1,1))
y.train <- rescale(as.matrix(wtloss$Weight), to=c(-1,1))
#training the ELM
trained.elm <- Elm.train(x.train,y.train,Number.hn =5)
#rescaling back the elm outputs
elm.fit.values <- rescale(trained.elm$predictionTrain,to= range(as.matrix(wtloss$Weight)),from=c(-1,1))
RMSE(wtloss$Weight,elm.fit.values)
oldpar <- par(mar = c(5.1, 4.1, 4.1, 4.1))
plot(wtloss$Days, wtloss$Weight, type = "p", ylab = "Weight (kg)",main="Weight Reduction")
lines(wtloss$Days, elm.fit.values,col=2,type='b')
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.