linear_baseline: Signal baseline

In phisanti/MyFunctions: My Functions

Description

Function to find the reference baseline in a fluorescence traceplot. The algorithm operates as follows: - First, it searches for the minimum value within the time series. - Second, removes $k$-neightbours before and after. - Third repeats $n$-times. - Fourth, fit a regression line to the filtered points.

The difference between the linear base liner and elastic baseliner is the regression method during the last step of the algorithm. In hte linear baseline uses the lm method while the elastic baseline relieas on a gam model.

Usage

linear_baseline(x, n_points = 20L, k = 5L, filter_lows = TRUE)

Arguments

x	'numeric': time series to find the baseline.
n_points	'integer': number of points to calculate the baseline.
k	'integer': number of points to clear each iteration.
filter_lows	'logic': whether or do a linear fitting to remove values above the trend.

Value

A list with three items: The estimated baseline, the adjusted residual value and the relative value.

Examples

x <- rnorm(500)
x <- x + 5* exp(-(1:500 - 200)^2/20) +
  5* exp(-(1:500 - 100)^2/20) +
  5* exp(-(1:500 - 400)^2/20)
plot(x, type = "b", pch = 16)

example <- linear_baseline(x)
example2 <- elastic_baseline(x)

lines(example$baseline, col = "red")
lines(example2$baseline, col = "blue")

phisanti/MyFunctions documentation built on Sept. 22, 2021, 7:37 p.m.