Adaptive EMA normalization for non-stationary EDA — standardise_gazepoint_adaptive

Applies dependency-light adaptive normalization using an exponential moving average center and robust local scale after IQR-based outlier screening. This preserves local dynamics more than whole-session z-scoring, but it is still a preprocessing transformation and not an emotion/stress classifier.

Usage

standardise_gazepoint_adaptive_ema(
  dat,
  signal_col = "GSR_US",
  group_cols = NULL,
  time_col = NULL,
  alpha = 0.05,
  iqr_multiplier = 1.5,
  suffix = "_adaptive_ema",
  center_suffix = "_ema_center",
  scale_suffix = "_ema_scale",
  min_scale = 1e-08,
  overwrite = FALSE
)

standardize_gazepoint_adaptive_ema(
  dat,
  signal_col = "GSR_US",
  group_cols = NULL,
  time_col = NULL,
  alpha = 0.05,
  iqr_multiplier = 1.5,
  suffix = "_adaptive_ema",
  center_suffix = "_ema_center",
  scale_suffix = "_ema_scale",
  min_scale = 1e-08,
  overwrite = FALSE
)

Arguments

dat: A data frame.
signal_col: Numeric signal column.
group_cols: Optional grouping columns.
time_col: Optional time column used to order rows within group.
alpha: EMA smoothing parameter in (0, 1].
iqr_multiplier: IQR multiplier for outlier screening.
suffix: Suffix for the normalized output column.
center_suffix: Suffix for the EMA center column.
scale_suffix: Suffix for the EMA scale column.
min_scale: Minimum scale used to avoid division by zero.
overwrite: Logical. If FALSE, existing output columns are protected.

Value

A data frame with adaptive normalized signal columns and attributes.