check.data: Function to check the dataset of individuals with (1) and...

In UncertainInterval: Uncertain Interval Methods for Three-Way Cut-Point Determination in Test Results

Description

Function to check the dataset of individuals with (1) and without (0) the targeted condition.

Usage

check.data(ref, test, model = c("kernel", "binormal", "ordinal", "none"))

Arguments

ref	The reference standard. A column in a data frame or a vector indicating the classification by the reference test. The reference standard must be coded either as 0 (absence of the condition) or 1 (presence of the condition)
test	The index test or test under evaluation. A column in a dataset or vector indicating the test results.
model	The model used for estimation. Default = 'kernel'. When model is kernel or binormal, the test data is checked whether the test has a sufficient number of different values (>= 20). When model is ordinal, the test data are checked whether they are ordinal or not. When model is 'none' the test data are only checked for missing data.

Details

The first check is whether ref and test have equal length. If not, checkdata is aborted with an error message. The second check is whether ref is coded solely with 0 and 1. If not, check.data is aborted and an error message is shown. The third check is whether ref and test have missing values. If true, list wise deletion is applied and a warning message is shown. The fourth check is whether test is continuous or not. If test has less than 20 different values, a warning message is shown. This test is omitted when ordinal = TRUE.

This function is called internally from every function that requires data. An external call is only useful to check warnings and errors.

Value

Either a valid dataset as data.frame with two variables ref and test or an error message.

Examples

set.seed(1)
ref=c(rep(0,500), rep(1,500))
test=c(rnorm(500,0,1), rnorm(500,1,1.2))
check.data(ref, test) # model = 'kernel'

UncertainInterval documentation built on March 3, 2021, 1:10 a.m.