MPFeatTrend_rollmean: MPFeatTrend_rollmean

In majpark21/TSexploreR: Explorative Tools for Time Series

Description

Smooth by rolling mean and perform linear regression. Rolling mean is extended at extremeties by linear extrapolation, see ?rollex. Aims at giving an estimation of long-term trend in multi-peak signals.

Usage

MPFeatTrend_rollmean(x, window.size, robust = FALSE)

Arguments

x	a numerical vector.
window.size	integer, width of window for rolling mean.
robust	logical. If TRUE perform robusts linear regression. Median-Based Linear Models are used, see ?mblm::mblm

Value

A list of 2: $model, the full linear model; $trend, slope of the linear model.

See Also

MPFeatTrend_extrema

Examples

width <- 30
x <- sim_phase_shifted_with_fixed_trend(n = 1, noise = 0, slope = 0.1)
# Rolling mean extended by linear extrapolation.
# This is also performed in MPFeatTrend_rollmean.
x.roll <- rollex(x$value, width)
plot(x$value, type = "b")
lines(x.roll, col ="red", lwd = 2, lty = "dashed")
x.trend <- MPFeatTrend_rollmean(x = x$value, window.size = width)
abline(x.trend$model, col = "blue", lwd = 2, lty = "dashed")

majpark21/TSexploreR documentation built on Oct. 16, 2019, 2:46 p.m.