Classification using normalized compression distance

Classification algorithms that use the NCD (Normalized Compression Distance) as a similarity metric are proposed. This way of measuring similarity allows either skipping the feature selection and feature extraction phases or extracting features in a more objective way than common feature extraction methods, and so makes the classification algorithms less biased This work consists of several classification experiments of images, voice samples and ECG signals using a weighted k nearest neighbor algorithm and support vector machines. Our results are comparative to those achieved by more complicated, parameter laden learning algorithms that do use feature selection and extraction.