Running Significance testing on LROC type data

[MPEngbersen]

Hi, I'm having trouble with the significance testing function of my LROC data. Using the latest GitHub version of the RJafroc package (v2.1.3) with R version 4.2.2.

I keep seeing the following errors with our own data but also with the data provided with the package:

library(RJafroc) packageVersion("RJafroc") [1] ‘2.1.3’ d <- DfFroc2Lroc(dataset04) p <- PlotEmpiricalOperatingCharacteristics(d, trts = list(1,2), rdrs = list( c(1,2), c(1,2)), opChType = "LROC") p$Plot St(d, FOM = 'ALROC', FPFValue = 0.2, method = 'OR') Error in AddTruthTableStr(dataset, dataset$descriptions$type, dataset$lesions$perCase) : data type must be ROC or FROC`

A different error message with the CRAN version (2.1.2) mentioned: Error in t[i, j, , 1] : incorrect number of dimensions

Am I missing something? Any help would be greatly appreciated.

[pwep]

Thanks for raising this issue. This appears to be due to a missing truthTableStr for the dataset.

• the DfFroc2Lroc function appears not to pass or recalculate a new truth table during the conversion
• the AddTruthTableStr function does not support an LROC type

These appear to have been relatively recent changes to the codebase, based on the comments @dpc10ster has added.

I will step through earlier versions of the package to try and catch where any breaking changes might have been made. I will add further updates to this issue as I progress.

[MPEngbersen]

Wonderful, thank you