constructTwoSampleModel: Constructing GP model with the specified kernels for two...

Description Usage Arguments Value Author(s) Examples

Description

Function for constructing the GP model with the specified kernels (and parameters) for two samples.

Usage

1
constructTwoSampleModel(x, y, v, kernelTypes, params = NULL)

Arguments

x

A list of one-column matrices which contain the input values for each series, i.e., time points for GP models. The values given in this vector are used in GP model, so if any transformation is needed, remember to transform them before constructing the model.

y

A list of one-column matrices which contain the observed values at the corresponding time points given in the x vector.

v

A list of one-column matrices which contain the fixed variances at the corresponding time points given in the x vector, with the corresponding observations given in the y vector.

kernelTypes

Character vector which contains the types of the kernels which will be used in the GP models. Kernel types: 'rbf', 'white', 'bias', 'fixedvariance'. Note that the lower bound for the length scale parameter of rbf kernel is set to the minimum distance between consecutive time points in order to mitigate potential overfitting problems.

params

Values of the kernel parameters in their transformed form. If not specified, default values are assigned.

Value

Return GP model constucted with the specified kernel settings.

Author(s)

Hande Topa, hande.topa@helsinki.fi; Antti Honkela, antti.honkela@helsinki.fi

Examples

1
2
3
4
5
6
7
x=list(as.matrix(seq(1,10)), as.matrix(seq(1,10)))
y=list(as.matrix(sin(x[[1]])), as.matrix(cos(x[[2]])))
v=list(as.matrix(runif(10,0,0.2)), as.matrix(runif(10,0,0.2)))
kernelTypes=c("rbf","white","fixedvariance")
model2=constructTwoSampleModel(x,y,v,kernelTypes)
model1=constructModel(rbind(x[[1]], x[[2]]), rbind(y[[1]],
+ y[[2]]), rbind(v[[1]], v[[2]]), kernelTypes)

PROBIC/GPrank documentation built on May 7, 2019, 11:53 p.m.