Computes the gradient numerically using finite differences. This is used
as a fallback when no analytical gradient is defined.
Usage
peak_model_gradient_numerical(model, x, params, eps = 1e-08)
Arguments
- model
A peak_model object.
- x
Numeric vector of x values.
- params
Named list of model parameters.
- eps
Step size for finite differences. Default is 1e-8.
Value
Matrix of partial derivatives with dimensions (length(x), n_params).
