The results below show a comparison of decoding accuracies for seven different classifier: correlation coefficient classifier (MCC, red), Euclidean distance classifier (EDist, gold), support vector machine (SVM, light green), regularized least squares (RLS, dark green), Poisson Naive Bayes (PNB, light blue), correlation coefficient classifier without first z-score normalizing the data (MCC no normalization, dark blue), and a raw dot product classifier (cyan). The upper three plots show the results from the 16 object stimulus set (combined data from both monkeys), and the lower three plots show the results from the 7 object stimulus set (for monkey 2 only). As can be seen, there is not a huge difference between the performance of different classifiers, although in some cases the MCC classifier without normalization seems to be performing a little worse than the other classifiers.
