Assessing Initial Results
Having completed several successful runs of the program, I now have data from trained models that I am reviewing and interpreting. This is not the dataset cited in the original work but it is related to the work and the programs were designed and tested with it in mind. The results from different computers are slightly different leading me to believe there are differences between the packages or versions used for underlying modules. When data was trained on a single machine multiple times, the results were the same so there is some other cause for the disparity. I have also realizing that the programs were never meant to test data from a set that they were not trained on. It is arguably more important for others to have access to the completed, trained model than the original data so I am revising the program to save the trained models for future use.