In conclusion, I trained the SVM data using a linear kernel, and each SVM was tested on the respective test data. The training and test datasets contained a roughly balanced number of positive and negative examples (0, 1). The confusion matrices (displayed in purple and pink units) can be found in the output above.
Train different SVM models using polynomial kernels with varying degrees (D = 1, 2, 3, 4, 5). Report the performance of these SVMs on both training and test datasets in terms of accuracy and AUC.