parallelSVM (version 0.1-9)

parallelSVM-package: Parallel-voting version of Support-Vector-Machine

Description

By sampling your data, running the Support-Vector-Machine algorithm on these samples in parallel on your own machine and letting your models vote on a prediction, we return much faster predictions than the regular Support-Vector-Machine and possibly even more accurate predictions.

Details

Package:

parallelSVM

Type:

Package

Version:

1.0

Date:

2015-02-09

License:

GPL-2

This package consists of two main functions: parallelSVM A function which allows you to create multiple Support-Vector-Machine models: one for each core you provide. It returns a list of Support-Vector-Machine models. predict: An extension of the predict function, which uses the prediction of each Support-Vector-Machine model. When probability is TRUE, it returns the average of all predictions, otherwise it returns the class most models agree upon.

See Also

This package can be regarded as a parallel extension of svm

Examples

## Not run: 
# # Create your data
# data(iris)
# x <- subset(iris, select = -Species)
# y <- iris$Species
# 
# # Create a model
# model <- parallelSVM(x, y)
# 
# # Get prediction
# predictions <- predict(model, x)
# 
# # Check the quality
# table(predictions,y)
# ## End(Not run)