Samples for ccdc

All samples were assumed to be correct when they were detected by all four basic algorithms described above at a consistent time (±1 year). False samples were identified as samples for which the results of the algorithms were not consistent with the latest CCDC breakpoints. True and false samples numbering 3,850 and 3,189, respectively were used to train and test the RF for identifying CCDC breakpoints In the dataset, the code of “presence” presents the true and false change for the sample. 0 presents the false change, and 1 presents the true change. The code of year presents the year of true change.