tsEvaTransformSeriesToStationaryTrendAndChangepts {RtsEva}

Transform Time Series to Stationary Trend and Change Points

Description

This function takes a time series and transforms it into a stationary trend series by removing the trend component and detecting change points. It computes the slowly varying standard deviation and normalizes the stationary series before detecting step changes. It also calculates the error on the trend and standard deviation.

Usage

tsEvaTransformSeriesToStationaryTrendAndChangepts(
  timeStamps,
  series,
  timeWindow
)

Arguments

timeStamps

A vector of time stamps corresponding to the data points in the series.

series

The original time series data.

timeWindow

The size of the time window used for detrending.

Value

A list containing the following elements:

runningStatsMulteplicity

The running statistics multiplicity.

stationarySeries

The transformed stationary series.

trendSeries

The trend series.

trendonlySeries

The trend series without the stationary component.

ChpointsSeries2

The trend component of the change points.

changePoints

The detected change points.

trendSeriesNonSeasonal

The trend series without the seasonal component.

trendError

The error on the trend.

stdDevSeries

The slowly varying standard deviation series.

stdDevSeriesNonStep

The slowly varying standard deviation series without step changes.

stdDevError

The error on the standard deviation.

timeStamps

The time stamps.

nonStatSeries

The original non-stationary series.

statSer3Mom

The running mean of the third moment of the stationary series.

statSer4Mom

The running mean of the fourth moment of the stationary series.

Examples

timeAndSeries <- ArdecheStMartin
timeStamps <- ArdecheStMartin[,1]
series <- ArdecheStMartin[,2]
#select only the 5 latest years
yrs <- as.integer(format(timeStamps, "%Y"))
tokeep <- which(yrs>=2015)
timeStamps <- timeStamps[tokeep]
series <- series[tokeep]
timeWindow <- 365 # 1 year
percentile <- 90
result <- tsEvaTransformSeriesToStationaryTrendAndChangepts(timeStamps,
series, timeWindow)
plot(result$trendSeries)