Skip to contents

Normalize confusion inputs to portable frequency records

Source: R/draw_confusion.R
confusion_input.Rd

Retains missing-label frequencies so omission is visible after serialization. Matrix dimnames carry class identity; columns are matched by name, never by an assumed diagonal position. No model dependency is needed for data inputs.

Usage

confusion_input(x, y = NULL, classes = NULL)

Arguments

x

Matrix, table, data frame, list of matrices, or vector: Counts or labels.

y

Optional vector: Predicted labels paired with reference labels in x.

classes

Optional Character vector: Explicit class order.

Value

Data frame with reference, predicted, n, and optionally panel columns.

Examples

confusion_input(factor(c("yes", "no")), factor(c("yes", "yes")))
#>   reference predicted n
#> 1        no        no 0
#> 2       yes        no 0
#> 3        no       yes 1
#> 4       yes       yes 1