i The k value entered. The number of folds used to estimate the performance of the model.

i The number of folds that successfully build a model.

i Each of the folds could be trained is considered significantly enriched if the p-value for predicting at least as many genes correctly is < 0.05 under the hypergeometric distribution.

i The times that the average performance improves over the random performance.

i Accuracy score considering known genes in the knowledge graph.

i Area under the receiver-operator characteristic curve for the trained model.

i F1 score considering known genes in the knowledge graph.

i Youden’s J statistic for the model, used as the training objective function.