| Title: | Extreme Machine Learning (ELM) |
| Version: | 1.0 |
| Author: | Alessio Petrozziello [aut, cre] |
| Maintainer: | Alessio Petrozziello <alessio.petrozziello@port.ac.uk> |
| Description: | Training and prediction functions are provided for the Extreme Learning Machine algorithm (ELM). The ELM use a Single Hidden Layer Feedforward Neural Network (SLFN) with random generated weights and no gradient-based backpropagation. The training time is very short and the online version allows to update the model using small chunk of the training set at each iteration. The only parameter to tune is the hidden layer size and the learning function. |
| Depends: | R (≥ 3.2.2) |
| License: | GPL-2 | GPL-3 |
| LazyData: | true |
| RoxygenNote: | 5.0.1 |
| NeedsCompilation: | no |
| Packaged: | 2015-11-28 11:19:30 UTC; Alessio |
| Repository: | CRAN |
| Date/Publication: | 2015-11-28 14:53:50 |
Trains an extreme learning machine with random weights
Description
Trains an extreme learning machine with random weights
Usage
OSelm_train.formula(formula, data, Elm_type, nHiddenNeurons, ActivationFunction,
N0, Block)
Arguments
formula |
a symbolic description of the model to be fitted. |
data |
training data frame containing the variables specified in formula. |
Elm_type |
select if the ELM must perform a "regression" or "classification" |
|
number of neurons in the hidden layer | |
ActivationFunction |
"rbf" for radial basis function with Gaussian kernels , "sig" for sigmoidal fucntion, "sin" for sine function, "hardlim" for hard limit function |
N0 |
size of the first block to be processed |
Block |
size of each chunk to be processed at each step |
Value
returns all the parameters used in the function, the weight matrix, the labels for the classification, the number of classes found, the bias, the beta activation function and the accuracy on the trainingset
Trains an online sequential extreme learning machine with random weights
Description
Trains an online sequential extreme learning machine with random weights
Usage
OSelm_training(p, y, Elm_Type, nHiddenNeurons, ActivationFunction, N0, Block)
Arguments
p |
dataset used to perform the training of the model |
y |
classes vector for classiication or regressors for regression |
Elm_Type |
select if the ELM must perform a "regression" or "classification" |
|
number of neurons in the hidden layer | |
ActivationFunction |
"rbf" for radial basis function with Gaussian kernels , "sig" for sigmoidal fucntion, "sin" for sine function, "hardlim" for hard limit function |
N0 |
size of the first block to be processed |
Block |
size of each chunk to be processed at each step |
Value
returns all the parameters used in the function, the weight matrix, the labels for the classification, the number of classes found, the bias, the beta activation function and the accuracy on the trainingset
References
[1] N.-Y. Liang, G.-B. Huang, P. Saratchandran, and N. Sundararajan, 'A Fast and Accurate On-line Sequential Learning Algorithm for Feedforward Networks' IEEE Transactions on Neural Networks, vol. 17, no. 6, pp. 1411-1423, 2006
Examples
x = runif(100, 0, 50)
y = sqrt(x)
train = data.frame(y,x)
train = data.frame(preProcess(train))
OSelm_train.formula(y~x, train, "regression", 100, "hardlim", 10, 10)
Pre processing function for the training and test data set. Each numeric variable is standardized between -1 and 1 and each categorical variable is coded with a dummy coding.
Description
Pre processing function for the training and test data set. Each numeric variable is standardized between -1 and 1 and each categorical variable is coded with a dummy coding.
Usage
preProcess(data)
Arguments
data |
to be preprocesses |
Value
return the pre processed dataset
Prediction function for the ELM model generated with the elm_training() function
Description
Prediction function for the ELM model generated with the elm_training() function
Usage
predict_elm(model, test)
Arguments
model |
the output of the elm_training() function |
test |
dataset used to perform the testing of the model, the first column must be the column to be fitted for the regression or the labels for the classification |
Value
returns the accuracy on the testset
References
[1] N.-Y. Liang, G.-B. Huang, P. Saratchandran, and N. Sundararajan, "A Fast and Accurate On-line Sequential Learning Algorithm for Feedforward Networks" IEEE Transactions on Neural Networks, vol. 17, no. 6, pp. 1411-1423, 2006
Examples
x = runif(100, 0, 50)
y = sqrt(x)
train = data.frame(y,x)
train = data.frame(preProcess(train))
model = OSelm_train.formula(y~x, train, "regression", 100, "hardlim", 10, 10)
#' x = runif(100, 0, 50)
y = sqrt(x)
test = data.frame(y,x)
test = data.frame(preProcess(train))
accuracy = predict_elm(model, test)