Finds optimal parameters for a set of peak models by minimizing the sum of squared residuals between the observed and fitted values.
Usage
optimize_deconvolution(
x,
y,
models,
init_params,
optimizer = "auto",
max_iter = 1000L,
tol = 1e-06,
constrain_positions = TRUE,
...
)Arguments
- x
Numeric vector of x-axis values (e.g., retention time, wavelength).
- y
Numeric vector of observed y-axis values.
- models
List of
peak_modelobjects, one per peak.- init_params
List of initial parameter lists, one per peak.
- optimizer
Optimization method:
"auto","lbfgsb","multistart", or"nelder_mead".- max_iter
Maximum number of iterations.
- tol
Convergence tolerance.
- constrain_positions
Logical. If
TRUE, fitted centers are nondecreasing in the rank of their starting centers and remain within each model's center bounds. Tied final centers are allowed. Duplicate starting centers, models without exactly one center parameter, and incompatible bounds are rejected. Optimization is local and does not guarantee a global minimum.- ...
Additional arguments passed to specific optimizers.
Value
A list containing:
parameters: List of optimized parameter listsfitted_values: Numeric vector of fitted y valuesresiduals: Numeric vector of residualsconvergence: Logical indicating convergencen_iterations: Number of iterations usedfinal_value: Final objective function value (SSE)optimizer: Name of optimizer usedelapsed_time: Optimization time in seconds
See also
Other peak-deconvolution:
add_param_jitter(),
assess_deconv_quality(),
check_quality_gates(),
initialize_peak_params()
Examples
# Create synthetic data with two overlapping Gaussian peaks
x <- seq(0, 20, by = 0.1)
true_y <- 1.5 * exp(-0.5 * ((x - 8) / 1)^2) +
0.8 * exp(-0.5 * ((x - 12) / 1.5)^2)
y <- true_y + rnorm(length(x), sd = 0.05)
# Set up models and initial guesses
models <- list(gaussian_peak_model(), gaussian_peak_model())
init_params <- list(
list(height = 1.2, center = 7.5, width = 1.2),
list(height = 0.6, center = 12.5, width = 1.8)
)
# Optimize
result <- optimize_deconvolution(x, y, models, init_params)
print(result$parameters)
#> [[1]]
#> [[1]]$height
#> [1] 1.495453
#>
#> [[1]]$center
#> [1] 8.000482
#>
#> [[1]]$width
#> [1] 1.00457
#>
#>
#> [[2]]
#> [[2]]$height
#> [1] 0.7901823
#>
#> [[2]]$center
#> [1] 12.00336
#>
#> [[2]]$width
#> [1] 1.516645
#>
#>