Fig. 2
From: An SVM-based method for assessment of transcription factor-DNA complex models
Performance of the SVM model by random splitting of the training set. The test was done by randomly splitting the original training set into training (106 complexes) and testing set (35 complexes). a Distribution of the Matthews correlation coefficients of 200 independent SVM models on the training and testing datasets using the radial basis functions (RBF) kernel. b Distribution of the accuracy of the SVM model compared to the accuracy from the orientation potential and DDNA3 potential