interpolate: Time Series Interpolation

In hallucigenia-sparsa/seqtime: Time Series Analysis of Sequencing Data

Time Series Interpolation

Description

Wrapper function to interpolate a time series. If groups are specified, each group is processed separately. The time points of each group are supposed to be in chronological order.

Usage

interpolate(x, interval = NA, time.index = NA, time.vector = c(),
  method = "stineman", groups = c(), negzero = FALSE)

Arguments

x	the time series matrix, rows are objects and columns are time points
interval	the target intervals to be present after interpolation
time.index	the row index holding time points
time.vector	the vector holding time points
method	fmm, periodic, natural, hyman or monoH.FC (spline), scaledstineman, stineman or parabola (stinterp)
groups	a vector with group assignments and as many entries as there are samples
negzero	set negative values (can be introduced by interpolation) to zero

Details

If no interval is provided, the intervals in the time vector are computed and the most frequent one is taken as the interval. The default interpolation method is stineman. Note that interpolation can introduce negative values in the abundance matrix.

Value

interpolated time series

Examples

  data("david_stoolA_otus")
  data("david_stoolA_metadata")
  days=david_stoolA_metadata[1,]
  sorted=sort(apply(david_stoolA_otus,1,sum),decreasing=TRUE,index.return=TRUE)
  davida.top=david_stoolA_otus[sorted$ix[1:100],]
  tsplot(interpolate(davida.top,time.vector=days))

hallucigenia-sparsa/seqtime documentation built on Jan. 9, 2023, 11:53 p.m.