4.1 Classification results

Overall results

None of the classifiers were able to correctly classify monocytes. The SVM classifiers classified basophils and neutrophils with the highest accuracy, and the neural networks classified eosinophils with the highest accuracy.

Svm 1 vs. 1

Figure 1. Confusion matrix for 1 Vs. 1 classification on test data. 1 = basophil,2 = eosinophil, 3 = lymphocyte,4 = monocyte, 5 = neutrophil.

Svm one vs. all

Figure 2. Confusion matrix for 1 Vs. All classification on test data.

Neural network

Figure 3. Confusion matrices for the training, validation, and test data. Results vary from class to class but are overall low. Like SVM, the neural network was not able to identify monocytes at all.

Figure 4. The Receiver Operating Characteristic (ROC) curve is a plot of the true positive rate vs. the false positive rate with varying threshold. A perfect test would show points in the upper-left corner