Skip to main content

Table 4 Accuracy by class for neural networks generated using one dataset as the training set and the other dataset for test data.

From: Prediction of peptides observable by mass spectrometry applied at the experimental set level

Class True positive rate False positive rate Precision Recall ROC Area
Avian Bursal Dataset training set, Hodgkins Lymphoma test set
Not observed 0.71 0.46 0.61 0.71 0.70
Observed 0.54 0.29 0.66 0.54 0.70
Hodgkin's Lymphoma Model Dataset training set, Avian Bursa test set
Not observed 0.81 0.41 0.81 0.73 0.73
Observed 0.59 0.19 0.59 0.66 0.73