predictTimecourses: Prediction of periodic time courses.

In MoPS: MoPS - Model-based Periodicity Screening

Description

Function that predicts periodic time courses using parameters identified by fit.periodic().

Usage

predictTimecourses(res.fits)

Arguments

res.fits

List object returned by fit.periodic().

Details

This function takes as input the result list from MoPS function fit.periodic() and creates a list of best fitting time courses. The input list also contains information about the screening parameters, which is used in the generation of predicted time courses.

Value

a numeric matrix containing the predicted values. The number of rows equals the number of rows of the original data matrix, the number of columns equals the number of screened phases.

Author(s)

Philipp Eser, Achim Tresch

Examples

x = seq(0,40,by=1) # time points

## create 10 periodic time series with added noise
mat.p = matrix(rep(x,10),nrow=10,ncol=length(x),byrow=TRUE)
y = -seq(1:10)
mat.p = apply(mat.p,2,function(x){
	y = sin(pi*(x/41*6)+y)+rnorm(length(x),sd=1)
	})

## add 10 non-periodic noisy time series
mat.nonP = matrix(rep(x,10),nrow=10,ncol=length(x),byrow=TRUE)
mat.nonP = apply(mat.nonP,2,function(x){
	y = rnorm(length(x),sd=1)
	})
	
mat = rbind(mat.p,mat.nonP)

res = fit.periodic(mat,phi=seq(0,20,1),lambda=seq(1,20,1))
time.courses = predictTimecourses(res)

plot(mat[1,],type="l",main="",xlab="",ylab="")
points(time.courses[1,],type="l",col="limegreen",lwd=2)

MoPS documentation built on April 28, 2020, 8:47 p.m.