The level of dexterity of myoelectric hand prostheses depends to large extent on the feature representation and subsequent classification of surface electromyography signals. This work presents a comparison of various feature extraction and classification methods on a large-scale surface electromyography database containing 52 different hand movements obtained from 27 subjects. Results indicate that simple feature representations as Mean Absolute Value and Waveform Length can achieve similar performance to the computationally more demanding marginal Discrete Wavelet Transform. With respect to classifiers, the Support Vector Machine was found to be the only method that consistently achieved top performance in combination with each feature extraction method.