Your English writing platform
Discover LudwigExact(4)
Comparison between different feature sets and different methods was mainly performed based on the AUC and ACC values.
For the solution in Eq. 6, a hard margin is constructed resulting in selection of only the most important base kernel, which eventually defeats the purpose of finding interaction between different feature sets.
The p value between different feature sets are plotted as bridge and stars, where two stars means p < 0.05 after both Bonferroni and FDR corrections, and one star means p < 0.05 only after FDR correction.
Fig. 1 Comparison between different feature sets of the four classical machine learning methods, based on mean AUCs and mean ACCs of the 10 times 10-fold cross-validation.
Similar(56)
While differences between the models were very small for cross-validation on the training set, some of them are emphasized on the test set, in particular between the different feature sets.
A reason for this was the focus on evaluating our feature set against the same data, only then we can make a comparison between two feature sets or a set of different feature sets.
Figure 4 Performance on different feature sets (soft case).
We then proceeded to compare three different feature sets.
Figure 3 Model performances on the different feature sets.
Comparison of prediction accuracy for different feature sets including random features.
In some hard cases for different feature sets (not shown here), especially for small feature sets with predominantly features that are not able to discriminate between classes, C5 as well as SVMs are performing as good as a majority class classifier since they predict everything as non-binding.
Write better and faster with AI suggestions while staying true to your unique style.
Since I tried Ludwig back in 2017, I have been constantly using it in both editing and translation. Ever since, I suggest it to my translators at ProSciEditing.

Justyna Jupowicz-Kozak
CEO of Professional Science Editing for Scientists @ prosciediting.com