Explain a model
explain.RdThis function is a generic method that dispatches to the appropriate explain method based on the class of the model.
Usage
explain(
model,
data = NULL,
model_type = NULL,
baseline_hazard = NULL,
labtrans = NULL,
time_bins = NULL,
preprocess_fun = NULL,
postprocess_fun = NULL,
predict_fun = NULL
)
# S3 method for class 'nn_module'
explain(
model,
data,
model_type,
baseline_hazard = NULL,
labtrans = NULL,
time_bins = NULL,
preprocess_fun = NULL,
postprocess_fun = NULL,
predict_fun = NULL
)
# S3 method for class 'extracted_survivalmodels_coxtime'
explain(model, data, ...)
# S3 method for class 'extracted_survivalmodels_deephit'
explain(model, data, ...)
# S3 method for class 'extracted_survivalmodels_deepsurv'
explain(model, data, ...)Arguments
- model
A model object.
- data
A data frame or matrix of data to explain the model.
- model_type
A string specifying the type of the survival model. Possible values are "coxtime", "deephit", or "deepsurv".
- baseline_hazard
A data frame containing the baseline hazard. It should have two columns: "time" and "hazard". This is only used for "coxtime" and "deepsurv" models.
- labtrans
A list containing the transformation functions for the time variable. It should have two elements: "transform" and "transform_inv". This is highly experimental and not yet fully supported.
- time_bins
A numeric vector specifying the time bins for the "deephit" model, e.g.,
c(0, 1, 2, 3).- preprocess_fun
A function to preprocess the data before making predictions, e.g., adding a time variable for a
coxtimemodel. This argument is highly experimental and the default values should work for most cases.- postprocess_fun
A function to postprocess the predictions after making them. This argument is highly experimental and the default values should work.
- predict_fun
A function that can be used to make predictions from the model. If
NULL, the predict method of the model will be used.- ...
Unused arguments.