Prediction Model Validation and Updating

Evaluate the predictive performance of an existing (i.e. previously developed) prediction/ prognostic model given relevant information about the existing prediction model (e.g. coefficients) and a new dataset. Provides a range of model updating methods that help tailor the existing model to the new dataset; see Su et al. (2018) . Techniques to aggregate multiple existing prediction models on the new data are also provided; see Debray et al. (2014) and Martin et al. (2018) ).

predRupdate

The goal of predRupdate is to provide a suite of functions for validating a existing (i.e. previously developed) prediction/ prognostic model, and for applying model updating methods to said model, according to an available dataset.

Installation

The package can be installed from CRAN as follows:

install.packages("predRupdate")

Development version

You can install the development version of predRupdate from GitHub with::

# install.packages("devtools")
devtools::install_github("GlenMartin31/predRupdate")

Example

One main use of this package is to externally validate an existing (previously developed) prediction model. This can be achieved with the following code:

# create a data.frame of the model coefficients, with columns being variables
coefs_table <- data.frame("Intercept" = -3.4,
                          "SexM" = 0.306,
                          "Smoking_Status" = 0.628,
                          "Diabetes" = 0.499,
                          "Creatinine" = 0.538)

#pass this into pred_input_info()
Existing_Logistic_Model <- pred_input_info(model_type = "logistic",
                                           model_info = coefs_table)
summary(Existing_Logistic_Model)

#validate this model against an available dataset
pred_validate(x = Existing_Logistic_Model,
              new_data = SYNPM$ValidationData,
              binary_outcome = "Y")