An Empirical Evaluation of Machine Learning Algorithms for Image Classification

Image classification is an important aspect that needs techniques which can better predict or classify images as they become larger and complex to solve. Thus, the demand for research to find advanced algorithms and tools to solve problems experienced in classification, has shown great increase in interest over the years. The contribution of this paper is the evaluation of four machine learning techniques [multilayer perceptron (MLP), random forests (RF), k-Nearest Neighbor (k-NN), and the Naive Bayes (NB)] in terms of classifying images. To this end, three industrial datasets are utilized against four performance measures (namely, precision, receiver operating characteristics, root mean squared error and mean absolute error). Experimental results show RF achieving higher accuracy while the NBC exhibits the worst performance.