Blockwise diagnostic for duration misspecification — diagnostic_duration

Compare the observed number of switches in a binary long-step feature within contiguous blocks to the predictive distribution under an HMM and an HSMM-like diagnostic alternative.

Usage

diagnostic_duration_blockwise(
  data,
  null_parameters,
  dwell_mean,
  dwell_size,
  step_threshold,
  block_size = 30L,
  n_simulations = 1000L,
  pseudo_count = 1,
  alpha = 0.05,
  diagnostic_name = "duration_blockwise",
  metadata = list()
)

Arguments

data: Movement data with step_length and turning_angle.
null_parameters: Null HMM parameters.
dwell_mean, dwell_size: HSMM-like dwell-time parameters for the diagnostic alternative.
step_threshold: Positive threshold defining long steps.
block_size: Number of observations per block.
n_simulations: Number of simulations for the HSMM block feature law.
pseudo_count: Non-negative pseudo-count for the simulated alternative feature distribution.
alpha: Monitoring level.
diagnostic_name: Diagnostic name.
metadata: Optional metadata list.

Value

A predictive_e_diagnostic object with block-level features.