Smart Parameter Initialization for Peak Deconvolution — initialize_peak_params • measure

Initializes peak model parameters using actual peak properties from the data rather than naive guesses, improving optimization convergence.

Usage

initialize_peak_params(
  x,
  y,
  n_peaks,
  models,
  peak_indices = NULL,
  smooth = TRUE,
  smooth_span = 0.05
)

Arguments

x: Numeric vector of x-axis values.
y: Numeric vector of y-axis values.
n_peaks: Number of peaks to initialize.
models: List of peak_model objects (one per peak).
peak_indices: Optional integer vector of peak indices (if already known).
smooth: Logical. If TRUE, smooth data before peak detection.
smooth_span: Smoothing span for LOESS (if smooth = TRUE).

Value

List of initialized parameter lists (one per peak).

See also

Other peak-deconvolution: add_param_jitter(), assess_deconv_quality(), check_quality_gates(), optimize_deconvolution()

Examples

# Create synthetic data with two peaks
x <- seq(0, 20, by = 0.1)
y <- 1.5 * exp(-0.5 * ((x - 8) / 1)^2) +
  0.8 * exp(-0.5 * ((x - 12) / 1.5)^2)

models <- list(gaussian_peak_model(), gaussian_peak_model())
init_params <- initialize_peak_params(x, y, n_peaks = 2, models = models)