Skip to main content

Table 1 The 10-CV training performance of the pre-training models and their prediction performances on test sets

From: Sequence-based prediction of protein protein interaction using a deep-learning algorithm

Code Sen. Spe. Pre. Acc. Test set acc. NR-test set acc.
AC 0.9806 0.9588 0.9581 0.9695 0.9682 0.9591
CT 0.9542 0.9367 0.9357 0.9452 0.9447 0.9312
  1. Column 2–5 represent the results of 10-cv with standard deviations ranged from 0.001 to 0.003
  2. Test set acc.: prediction accuracy for the hold-out test set
  3. NR-test set acc.: prediction accuracy for the NR-test set