smoothLoess: Smooth time series using loess
In johanez/probaV: Utils for time-series analysis of Proba-V data

Description Usage Arguments Details Value

Temporal outliers are marked or removed based on their distanced from a smoothing model derived with loess.

1 2	smoothLoess(tsx, QC_good = NULL, dates = NULL, threshold = c(-50, Inf), res_type = c("distance", "sd_distance", "all", "filled", "omit", "QC"), ...)

`tsx`	ts or numeric vector. Time series data.
`QC_good`	Integer or Logical vector. Optional time series of ts quality: 0=bad, 1=good
`dates`	Dates or Numeric. Vector of observation times. Overwrites index of tsx.
`threshold`	Numeric vector of length 2. Upper and lower threshold for non-outlier distances from smoothed time series.
`res_type`	Character. Determins the returned object , see details.
`...`	additional arguments to `loess`

"distance" gives the distance, "sd_distance" the distance devided by the sd, "filled", a ts with outleirs replaced by smoothed values, "omit" the input ts with outliers removed, "QC" the status of observations: 0=Missing/band input QC, 1=good, 2=temporal outlier. "all" returns a data.frame with all these vectors.

a ts, numeric vector or data.frame, see details

johanez/probaV documentation built on May 19, 2019, 3:01 p.m.