The classification of the electrocardiogram (ECG) into different pathophysiological disease categories is a complex pattern recognition task. This study compares the classification performance of feature sets formed from the discrete-wavelet-transform coefficients of different mother wavelets. Fifteen feature sets are calculated from three Daubechies wavelets, with the decomposition level varied between three and seven. Classification performance is optimized by using automatic feature selection and by combining classifications of multi-beat ECG information. Results show that the overall classification performance of the different feature sets was 71.6-74.2% and that the wavelet order and level had little influence on the overall performance. All quoted results are obtained from 10 runs of 10-fold cross-validation. Analysis of the automatically chosen features reveals that time-frequency bands approximately the QRS onset and the T-wave is consistently selected.