adjust-probability-calibration.md
In tailor: Iterative Steps for Postprocessing Model Predictions

adjustment printing

Code
  adjust_probability_calibration(tailor())
Message

  -- tailor ----------------------------------------------------------------------
  A binary postprocessor with 1 adjustment:

  * Re-calibrate classification probabilities using method.

Code
  adjust_probability_calibration(tailor(), method = "logistic")
Message

  -- tailor ----------------------------------------------------------------------
  A binary postprocessor with 1 adjustment:

  * Re-calibrate classification probabilities using logistic method.

Code
  adjust_probability_calibration(tailor(), method = hardhat::tune())
Message

  -- tailor ----------------------------------------------------------------------
  A binary postprocessor with 1 adjustment:

  * Re-calibrate classification probabilities (method marked for optimization).

Code
  fit(adjust_probability_calibration(tailor()), d_bin_calibration, outcome = c(y),
  estimate = c(predicted), probabilities = c(a, b))
Message

  -- tailor ----------------------------------------------------------------------
  A binary postprocessor with 1 adjustment:

  * Re-calibrate classification probabilities using method. [trained]

errors informatively with bad input

Code
  adjust_probability_calibration(tailor(), "boop")
Condition
  Error in `adjust_probability_calibration()`:
  ! `method` must be one of "logistic", "multinomial", "beta", "isotonic", "isotonic_boot", or "none", not "boop".

Code
  adjust_probability_calibration(tailor(), "linear")
Condition
  Error in `adjust_probability_calibration()`:
  ! `method` must be one of "logistic", "multinomial", "beta", "isotonic", "isotonic_boot", or "none", not "linear".

tuning the calibration method

Code
  fit(tlr, d_bin_calibration, outcome = c(y), estimate = c(predicted),
  probabilities = c(a, b))
Condition
  Error in `fit()`:
  ! The calibration method cannot be a value of `tune()` at `fit()` time.

too few data

Code
  tlr_fit <- fit(tlr, d_bin_calibration[0, ], outcome = c(y), estimate = c(
    predicted), probabilities = c(a, b))
Condition
  Warning:
  The calibration data has 0 rows. There is not enough data for calibration so `method` is changed from "logistic" to "none".

Code
  tlr_fit <- fit(tlr, d_bin_calibration[1, ], outcome = c(y), estimate = c(
    predicted), probabilities = c(a, b))
Condition
  Warning:
  The calibration data has 1 row. There is not enough data for calibration so `method` is changed from "logistic" to "none".

passing arguments to adjust_probability_calibration

Code
  tlr_fit <- adjust_probability_calibration(tailor(), method = "logistic", FALSE)
Condition
  Error in `adjust_probability_calibration()`:
  ! All calibration arguments passed to `...` should have names.

harden against calibration model failure

Code
  y_fit <- fit(tlr, d_y_calibration, outcome = c(truth), estimate = c(predicted),
  probabilities = c(Class1, Class2))
Condition
  Warning:
  glm.fit: algorithm did not converge
  Warning:
  glm.fit: algorithm did not converge
  Warning:
  The beta calibration failed. No calibration is applied.
  i Error in uniroot(function(mh) b * log(1 - mh) - a * log(mh) - inter, c(1e-16, : f() values at end points not of opposite sign