BCA (version 0.9-3)

score: Score a Database based on a Predictive Model

Description

Provides either an integer vector that contains the "desirability" rank of a case in a data set, the fitted probability of a desired response, or the fitted probability adjusted for the true response rate based on the fitted values of a predictive model.

Usage

rankScore(model, data, targLevel)
rawProbScore(model, data, targLevel)
adjProbScore(model, data, targLevel, trueResp)

Arguments

model

A character string containing the name of the model to use to score the database.

data

A data frame of the database to be scored. All the predictor variables of the model need to be amoung the variables of the data frame.

targLevel

The "desired" level of the y variable factor as a character string.

trueResp

The true "desired" response rate for the overall population of interest.

Value

An integer vector that indicates the rank order desirability (a value of 1 means most desirable) of the correponding case of the database being scored or a probability measure bounded between zero and one.

Details

Only binomial glm, binary rpart, and binary nnet models can be used as the basis of scoring a database.