Predict detection function from fitted SNR models

Unified prediction interface for SNR-based detection models. Returns fitted detection probabilities over a sequence of SNR values (or user-supplied newdata), optionally with 95% confidence intervals.

Supports:

• glm, gam, scam (mgcv/scam-style models)

• VGAM vglm / vgam (posbernoulli.t models)

Usage

predictDetFun(model, ...)

Arguments

model

Fitted model from: fitSNRdetectionFunc(), fitSNRvglm(), or fitSNRvgam()

newdata

Data.frame with column SNR. If NULL, a regular grid is generated from the model frame.

ci

Logical. If TRUE, compute 95% confidence intervals.

npoints

If newdata = NULL, number of SNR grid points.

nsim

Number of simulations for VGAM confidence intervals.

Value

A data.frame with:

• SNR

• fit

• lower (optional)

• upper (optional)