I'm looking through the paper trying to figure out what the median absolute error (MAE) or root mean square error (RMSE) of DA06 (or similar) would be if a constant, average model was used. Since DA06 all have ground truth values, if we took the average of all 60,000 ground truth grain boundary energy (GBE) values and used this as a constant-valued model, what would the RMSE and MAE be? This is related to Issue #7 that I just opened. My plan was to generate a ground truth file, but it doesn't look like it has GBE as one of the columns.
Update -- I feel a bit silly, but just noticed/remembered TJdata/trueE_30000.txt. I think this is all I need for now.
Hi Yufeng,
I'm looking through the paper trying to figure out what the median absolute error (MAE) or root mean square error (RMSE) of DA06 (or similar) would be if a constant, average model was used. Since DA06 all have ground truth values, if we took the average of all 60,000 ground truth grain boundary energy (GBE) values and used this as a constant-valued model, what would the RMSE and MAE be? This is related to Issue #7 that I just opened. My plan was to generate a ground truth file, but it doesn't look like it has GBE as one of the columns.
Update -- I feel a bit silly, but just noticed/remembered TJdata/trueE_30000.txt. I think this is all I need for now.
Sterling