prediction.Classification: Prediction from MKL model

In cwilso6/RMKL: Multiple Kernel Learning for Classification, Regression Problems, and Survival Settings

Description

This function creates gramm matrix for traning set baed upon several types of kernel and specified hyper paremeters. Matrix corresponds to similarity betwween each sample in the training set.

Usage

prediction.Classification(model, ktest, train.outcome)

Arguments

model	MKL model
ktest	Gramm matrix of training data and test data
train.outcome	Outcome for the training data

Value

yhat Predicted value for each test point

predicted Sign of yhat, which is the final predicted outcome

References

A Rakotomamonjy, FR Bach, S Canu, Y Grandvalet. Journal of Machine Learning Research 9 (Nov), 2491-2521.

Examples

library(kernlab)
library(caret)
data(benchmark.data)
example.data=benchmark.data[[1]]
training.samples=sample(1:dim(example.data)[1],floor(0.7*dim(example.data)[1]),replace=FALSE)
C=100
kernels=rep('radial',3)
degree=rep(0,3)
scale=rep(0,3)
sigma=c(0,2^seq(-3:0))
K=kernels.gen(example.data[,1:2], training.samples, kernels, degree, scale, sigma)
K.train=K$K.train
K.test=K$K.test
SEMKL.model=SEMKL.classification(K.train,example.data[training.samples,3], C)
predicted=prediction.Classification(SEMKL.model, K.test, example.data[training.samples,3])
confusionMatrix(factor(predicted$predict, levels=c(-1,1)),
                factor(example.data[-training.samples,3],levels=c(-1,1)))

cwilso6/RMKL documentation built on May 18, 2021, 9:58 a.m.