dataStateSwitch: Alternate Observations in a Data Frame over States

In gabrielodom/mvMonitoring: Multi-State Adaptive Dynamic Principal Component Analysis for Multivariate Process Monitoring

Alternate Observations in a Data Frame over States

Description

Split single-state process observations, apply multiple state projections, and combine these observations into a single data frame, arranged by process time or index.

Usage

dataStateSwitch(
  df,
  angles2 = list(yaw = 0, pitch = 90, roll = 30),
  scales2 = c(1, 0.5, 2),
  angles3 = list(yaw = 90, pitch = 0, roll = -30),
  scales3 = c(0.25, 0.1, 0.75)
)

Arguments

df	A data frame returned by processNOCdata() or faultSwitch().
angles2	Change the principal angles for State 2.
scales2	Change the principal scales for State 2.
angles3	Change the principal angles for State 3.
scales3	Change the principal scales for State 3.

Details

This function splits a process data frame by state, and rotates and scales the observations from states 2 and 3 by the scales and angles specified in the function arguments. After state-specific rotation and scaling, this function combines the observations back together and orders them by process time index. This function takes in data frame returned by processNOCdata() or faultSwitch(). This function calls rotateScale3D() and is called internally by mspProcessData().

Value

A data frame containing the time index, state, and feature values after state-specific rotation and scaling; this data frame also contains the other columns of df that aren't the feature values. This data frame has

dateTime -

a POSIX column of the time stamps for each observation

state -

column of state membership (1, 2, or 3)

x -

the process values for the first feature, corresponding to t + random error

y -

the process values for the second feature, corresponding to t ^ 2 - 3 * t + random error

z -

the process values for the third feature, corresponding to -t ^ 3 + 3 * t ^ 2 + random error

t -

the non-stationary and autocorrelated latent feature

err1 -

a Gaussian white noise vector

err2 -

a Gaussian white noise vector

err3 -

a Gaussian white noise vector

See Also

Calls: processNOCdata, faultSwitch, rotateScale3D. Called by: mspProcessData

Examples

nrml <- processNOCdata()
dataStateSwitch(nrml)

gabrielodom/mvMonitoring documentation built on Nov. 23, 2023, 6:39 p.m.