Skip to main content

Table 2 Some classification algorithms perform better than others, but all yield similar conclusions

From: Extrapolating histone marks across developmental stages, tissues, and species: an enhancer prediction case study

  Heart features Non-heart features
Random forest 0.85 0.72
Linear SVM 0.84 0.73
AdaBoost 0.82 0.70
Naive bayes 0.79 0.69
Decision tree 0.77 0.62
KNN (k=3) 0.74 0.66
  1. This table gives ROC AUCs (averaged over five cross-validation folds) for six common algorithms at distinguishing E11.5 heart enhancers from other enhancers based on marks from heart or non-heart tissues.