Please correct me if I am wrong, but I believe the jack-prop argument is inverted.
It is described as:
--jack-prop NUM Subsampling proportion for jackknife (default: 0.5)
In the tools.cpp, the code uses it as follows:
int total = floor((1.0 - Params::getInstance().jackknife_prop)*n);
if (total <= 0)
outError("Jackknife sample size is zero");
If jack-prop is 1 (defined as subsampling all the data [not subsampling]), then total would be zero and no sites would be selected (throwing an error). If jack-prop is 0 (defined as sampling no data), then total would be n and all sites would be selected.
I might be thinking about this backwards, but I believe jack-prop should be defined as:
--jack-prop NUM Deletion proportion for jackknife (default: 0.5)
Of course, this has no effect at the default value since 1 - 0.5 = 0.5.
Most importantly, thank you for maintaining the program and taking a look at this report.
Please correct me if I am wrong, but I believe the jack-prop argument is inverted.
It is described as:
--jack-prop NUM Subsampling proportion for jackknife (default: 0.5)In the tools.cpp, the code uses it as follows:
If
jack-propis 1 (defined as subsampling all the data [not subsampling]), then total would be zero and no sites would be selected (throwing an error). Ifjack-propis 0 (defined as sampling no data), then total would be n and all sites would be selected.I might be thinking about this backwards, but I believe jack-prop should be defined as:
--jack-prop NUM Deletion proportion for jackknife (default: 0.5)Of course, this has no effect at the default value since 1 - 0.5 = 0.5.
Most importantly, thank you for maintaining the program and taking a look at this report.