Figure 1
From: PoGO: Prediction of Gene Ontology terms for fungal proteins
Overview of the PoGO training and evaluation procedure. The PoGO classifier uses a Combiner configuration in which two base-level classifiers are trained on 45% of the training proteins and evaluated on another 45% of the training proteins. The remaining 10% of the training proteins are used to evaluate the meta-classifier. This configuration is repeated for each of the Gene Ontology terms trained in PoGO.