dddd: One reader and Multiple modality data

In BayesianFROC: FROC Analysis by Bayesian Approaches

Description

This is a subset of dd. For this dataset, the function fit_Bayesian_FROC() well works. So, even if the number of reader is one, my programm is available. Even if not available, I think it does not cause my model but my programming.

dddd$M

5 modalities

dddd$C

5 Confidence levels

dddd$Q

1 readers

Details

Model converged in 2019 Jun 21.

Contents of dddd

NL = 142 (Number of Lesions)

NI = 199 (Number of Images)

Contents:

Multiple readers and multiple modalities case, i.e., MRMC case

—————————————————————————————————

ModalityID	ReaderID	Confidence levels	No. of false alarms	No. of hits.
q	m	c	f	h
--------------	-------------	------------------------	-------------------	----------------
1	1	5	0	50
1	1	4	4	30
1	1	3	20	11
1	1	2	29	5
1	1	1	21	1
2	1	5	1	52
2	1	4	1	25
2	1	3	21	13
2	1	2	24	4
2	1	1	23	1
3	1	5	1	43
3	1	4	7	29
3	1	3	13	11
3	1	2	28	6
3	1	1	19	0
4	1	5	1	61
4	1	4	4	19
4	1	3	18	12
4	1	2	21	9
4	1	1	23	3
5	1	5	0	35
5	1	4	2	29
5	1	3	19	18
5	1	2	23	9
5	1	1	18	0

—————————————————————————————————

The reason why the author made this data dddd is it has only one reader. My program well works for more than two reader and more than two modality case. However, the only one modality or only one reader case is very special for programming perspective, and thus the author had to confirm whether my program well works in such cases. For this dataset, the function fit_Bayesian_FROC() well works. So, even if in a single reader case, my programm is available. Even if not available, I think it does not cause my model but my programming.

References

Example data of Jafroc software

See Also

dataList.Chakra.Web dataList.Chakra.Web.orderd dd

Examples

#========================================================================================
#                        Show data by table
#========================================================================================



                 viewdata(BayesianFROC::dddd)




#========================================================================================
#              make an object dddd from an object dd
#========================================================================================



           ddd  <-  data.frame(m=dd$m,q=dd$q,c=dd$c,h=dd$h,f=dd$f)

           dddd <-  ddd[ddd$q < 2,]  #  Reduce the dataset ddd, i.e., dd

ddd <- list(
           m=dddd$m,
           q=dddd$q,
           c=dddd$c,
           h=dddd$h,
           f=dddd$f,
           NL=142,
           NI=199, # 2020 April 6
           C=max(dddd$c),
           M=max(dddd$m),
           Q=max(dddd$q)
        )

          dddd <-ddd


#========================================================================================
#              Fit model to the object dddd
#========================================================================================
#  Unfortunately, R CMD check require running time to be less than 5 which is difficult
#  for rstan::sampling(), thus, we cannot run the following from roxygen2 example.
#
#
#     For Fitting, execute the following R code;
#
#
#

BayesianFROC documentation built on Jan. 23, 2022, 9:06 a.m.