Figure 7
From: A novel approach for protein subcellular location prediction using amino acid exposure
ROC curves from SVM classifications (winner-takes-all strategy) and ANN classifications that use as input the SVM values. For SVMs the ROC curves (dotted lines) were made by taking the best prediction from sets of SVMs (winner-takes-all strategy). Either best of four SVMs for each location category (red, green and cyan dotted curves indicating the different ranges used), or best of 12 SVMs (the combination of three SVM types is indicated with pipe signs indicating the vectors used; violet dotted curve). For ANNs the ROC curves (continuous lines) used just the ANN output. See text for details.