Comparative Study of Image Classification using Machine Learning Algorithms

Abstract

This study compared five common machine learning algorithms for performing classification included Support Vector Machines (SVM), K-Nearest Neighbor (KNN), Naïve Bayes (NB), Binary Decision Tree (BDT) and Discriminant Analysis (DA). AlexNet deep learning model was used to build these machine learning classifiers. The building classifiers were implemented and evaluated according to standard performance criteria of Accuracy (ACC), Precision (P), Sensitivity (S), Specificity (Spe) and Area Under the ROC Curve (AUC). The five methods were evaluated using 2608 histopathological images for head and neck cancer. The comparison was conducted using 2 times 10-fold cross validation. For each method, the pre-trained AlexNet network was used to extract features from the activation layer. The results illustrated that, there was no difference between the results of SVM and KNN. Both have the same and the higher accuracy than others were 99.98 %, whereas 99.81%, 97.32% and 93.68% for DA, BDT and NB, respectively. The present study shows that the SVM, KNN and DA are the best methods for classifying our dataset images.