Currently, the demo notebook only show results for one split of the data, so we don't know whether this performance is sensitive to how the data is split. We need to test this sensitivity and thus the robustness of the results achieved so far by doing cross validation on the classifiers tried so far.
Currently, the demo notebook only show results for one split of the data, so we don't know whether this performance is sensitive to how the data is split. We need to test this sensitivity and thus the robustness of the results achieved so far by doing cross validation on the classifiers tried so far.