File:Fpls-02-00043-g001.png

English: Outline of prediction engine construction and evaluation. The prediction engine was constructed and its performance was evaluated in two independent runs on the entire working data set, as indicated by the green and blue arrows respectively. (A) Assessment of prediction engine performance through fivefold cross-validation (CV) loop. In each fold of CV an optimized classifier is built on the training set and its performance is assessed on the test set. The resulting performance measure distribution of the optimized classifiers was used to assess the performance of the prediction engine. (B) The training procedure used during prediction engine construction and evaluation. The training of a classifier involved feature selection and estimation of the best Cost (C) and gamma (Γ) parameters using a grid search with fivefold CV loop. The best performing parameter pair was then used to train an optimized classifier.