clean_artifact: Automatically clean artifacts.
In PupilPre: Preprocessing Pupil Size Data

Description Usage Arguments Value References Examples

clean_artifact performs two stage and distributional automated clean-up of artifacts in the pupil and gaze coordinate data.

clean_artifact(data = data, MADWindow = 100, MADConstant = 2,
  MADPadding = c(200, 200), MahaConstant = 2, Method = "Robust",
  XandY = TRUE, Second = TRUE, MaxValueRun = 5, NAsAroundRun = c(2,
  2), LogFile = NULL)

`data`	A data frame object created from `ppl_select_recorded_eye`.
`MADWindow`	A numeric value specifying the window size (in msec) to use for the MAD calculation.
`MADConstant`	A numeric value specifying the constant (a multiplier for the third quartile) when determining MAD outlier status.
`MADPadding`	A numeric vector of length two containing values (in msec) to pad the identified artifact creating a window within which to operate the cleanup.
`MahaConstant`	A numeric value specifying the constant (a multiplier for the third quartile) when determining Mahalanobis outlier status.
`Method`	A character string ("Basic" or "Robust") indicating which method to use for the distance calculation. Basic is a standard Mahalanobis distance calculation based on covariance. Robust is also a Mahalanobis distance, however, it is based on Minimum Covariance Determinant (Rousseeuw and van Driessen, 1999) with reweighted covariance (Pison et al., 2002). For more details, see `covMcd`.
`XandY`	A logical value specifying whether to also use horizontal velocity and acceleration in outlier detection.
`Second`	A logical value specifying whether secondary cleaning should be applied.
`MaxValueRun`	A numeric value specifying the maximal run of existing values flanked by NAs that could be targeted for removal.
`NAsAroundRun`	A numeric vector of length two containing values (in number of subsequent NA) to be used to identify straggler runs of data that could be removed.
`LogFile`	A character string indicating the file name (with extension) of the log file to be created/written. The file keeps track of which events have been cleaned. We suggest "ArtifactCleanupLog.rds".

An object of type data table as described in tibble.

Rousseeuw, P. J. and van Driessen, K. (1999) A fast algorithm for the minimum covariance determinant estimator. Technometrics 41, 212–223.

Pison, G., Van Aelst, S., and Willems, G. (2002) Small Sample Corrections for LTS and MCD, Metrika 55, 111–123.

# Load example data
data("Pupilex4")

# Writing log file to temporary folder for the example
dat <- clean_artifact(Pupilex4, MADWindow = 100, MADConstant = 2,
                      MADPadding = c(200, 200), MahaConstant = 2,
                      Method = "Robust", XandY = TRUE, Second = TRUE,
                      MaxValueRun = 5, NAsAroundRun = c(2,2),
                      LogFile = paste0(tempdir(),"/ArtifactCleanupLog.rds"))

# Please see the vignettes for detailed example usage.
# vignette("PupilPre_Cleanup", package="PupilPre")