Skip to contents

Add a d2y_dx2 column to an existing prediction grid

Source: R/detection_limits.R
enrich_grid_with_d2y.Rd

Computes \(d^2(\log_{10} y) / d(\log_{10} x)^2\) from the log10_concentration and predicted_response columns already present on the grid, using non-uniform central differences.

Usage

enrich_grid_with_d2y(grid, is_log_response = TRUE)

Arguments

grid

Data frame with columns log10_concentration and predicted_response.

is_log_response

Logical; whether predicted_response is already on the \(\log_{10}\) scale (default TRUE).

Value

The input grid with an additional column d2y_dx2. Boundary points (first and last rows) receive NA.

Details

This function is designed to be called inside predict_grid_freq() or predict_grid_bayes() immediately after the grid data frame is constructed, adding the column in-place with zero additional model evaluations.

When is_log_response = TRUE, the predicted_response column is already on the \(\log_{10}\) scale, so d2y_dx2 is computed directly. When is_log_response = FALSE, the column is first \(\log_{10}\)-transformed before differencing.