process.zacl: Process Zio ZACL File

In etzkorn/postuR: Posture Classification and Activity Measures for Triaxial Accelerometers in ECG Patches.

Process Zio ZACL File

Description

Takes the raw Zio accelerometer file and processes it into adjacent intervals of length epoch.seconds.

Usage

process.zacl(
  data,
  epoch.seconds = 60,
  p = 0.95,
  k = 0.98,
  theta.star = 45,
  screen.nonwear = T,
  check.device.rotation = T,
  nonwear.window = 3 * 94 * 60,
  nonwear.tol = 10,
  nonwear.tol.upper = 0,
  minimum.wear.bout = 94 * 60 * 24,
  cluster = "none"
)

Arguments

data	Data frame containing the accelerometer data. Should include variables time, x, y, z.
epoch.seconds	Duration (seconds) of interval in which to summarise the data. Options include 10, 30, 60, 300, or "none". If epoch.seconds is "none", then data is processed at the original frequency of the data.#'
p	percentile for defining high-acceleration point. Used in estimation of upright orientation.
k	ratio cutoff for detecting removal/rotation/reattachment.
theta.star	Threshold for inclination for determining whether a point is recumbent or upright.
check.device.rotation	Boolean. Should device removal/rotation/reattachment be screened for?
nonwear.window	Window length (units of sampling points on raw data scale) representing minimum possible nonwear period. Passed to accelerometry::weartime. Defaults to 16920, about 3 hours of Zio wear time.
nonwear.tol	Number of allowed changes in window length to be considered nonwear. Defaults to 10.
nonwear.tol.upper	Maximum allowable sum of change in axis gravitational units in any window of non-wear. To be passed to accelerometry::weartime. Default of NULL is equivalent to zero.
minimum.wear.bout	Window length (units of sampling points on raw data scale) representing minimum possible wear period. Defaults to 135360, about 24 hours of Zio wear time.
cluster	Clustering algorithm to create posture groups. Either "meanShift", "ward", "centroid", or "none". Choosing "none" will dramatically reduce processing time.
check.nonwear	Boolean. Should non-wear be screened for?

Value

A data frame with epoch level activity summaries. For each time stamp in the resulting data frame, the summary corresponds to the following epoch.seconds (i.e. 22/9/19 10:50:00 corresponds to the time interval from 22/9/19 10:50:00 to 22/9/19 10:54:59 if epoch.seconds=300).

cluster	Which posture cluster does an epoch belong to?
down	Indicator for recumbent.
wear.bout	Integer index for period of consecutive wear epochs an epoch corresponds to.
theta	estimated inclination of the chest during the epoch.
mad	Mean absolute deviation, a measure of activity intensity during the epoch (milli-gravitational units).
x, y, z	Median accelerations measured along each axis during an epoch (gravitational units).

etzkorn/postuR documentation built on Nov. 15, 2024, 12:41 a.m.